JP2017182363A - Information processing device, information processing method and information processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information processing device, an information processing method and an information processing program capable of showing in real time a suggestion to a reason in which performance fluctuates and a suggestion as to how much influence the reason gives to the fluctuation of the performance.SOLUTION: The information processing device includes a first instruction part 11 for instructing an arbitrary explanatory variable, an acquisition part 12 for acquiring an information source corresponding to the arbitrary explanatory variable, an analysis part 13 for analyzing the information source, a first waveform generation part 14 for generating a first waveform on the basis of a result analyzed by the analysis part 13, a second instruction part 15 for instructing an arbitrary objective variable, a second waveform generation part 16 for generating a second waveform on the basis of relevance between a waveform of the arbitrary objective variable and the first waveform, and an output part 17 for outputting the second waveform to a display part.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an information processing apparatus, an information processing method, and an information processing program for processing information.

  For example, Patent Document 1 proposes a method and system apparatus for processing user history information based on a Web access log, a product / service purchase log, and the like, and predicting an item that may become the next trend of the user. Yes.

JP 2001-350875 A

  By the way, even if the method and apparatus disclosed in Patent Document 1 are applied, it is difficult to indicate the reason for the temporal fluctuation of the performance such as the sales of the product or the stock price. It is difficult to show how much it affects.

  For example, if you can show why a change in performance that a product sold or did not sell, and why it affects that change, marketing Useful in many situations.

  The present invention provides an information processing apparatus, an information processing method, and an information processing program capable of showing in real time an indication of the reason for the change in performance and how much the reason affects the change in the performance. For the purpose.

  To achieve the above object, an information processing apparatus according to an aspect of the present invention includes a first instruction unit that instructs an arbitrary explanatory variable, an acquisition unit that acquires an information source corresponding to the arbitrary explanatory variable, An analysis unit that analyzes an information source, a first waveform generation unit that generates a first waveform based on a result analyzed by the analysis unit, a second instruction unit that indicates an arbitrary objective variable, and the arbitrary purpose A second waveform generation unit configured to generate a second waveform based on a relationship between the variable waveform and the first waveform; and an output unit configured to output the second waveform to a display unit.

  An information processing method according to one aspect of the present invention includes a first instruction step for indicating an arbitrary explanatory variable, an acquisition step for acquiring an information source corresponding to the arbitrary explanatory variable, and an analysis step for analyzing the information source. A first waveform generating step for generating a first waveform based on a result analyzed by the analyzing step, a second indicating step for indicating an arbitrary objective variable, a waveform of the arbitrary objective variable, and the first waveform A second waveform generating step for generating a second waveform based on the relationship with the output, and an output step for outputting the second waveform to a display unit.

  An information processing program according to an aspect of the present invention includes a first instruction step for indicating an arbitrary explanatory variable, an acquisition step for acquiring an information source corresponding to the arbitrary explanatory variable, and an analysis step for analyzing the information source. A first waveform generating step for generating a first waveform based on a result analyzed by the analyzing step, a second indicating step for indicating an arbitrary objective variable, a waveform of the arbitrary objective variable, and the first waveform Is a program for realizing, by a computer, a second waveform generation step for generating a second waveform based on the relationship between the output and the output step for outputting the second waveform to a display unit.

  According to the present invention, it is possible to show in real time an indication for the reason why the performance has changed and an indication of how much the reason has an influence on the change in the performance.

It is a block diagram which shows the structure of information processing apparatus. It is a block diagram which shows the structure of an analysis part. It is a figure which shows an example of a 1st waveform typically. It is a figure which shows an example of a 2nd waveform typically. It is a figure which shows an example of a 1st operation screen. It is a figure which shows an example of a 2nd operation screen. It is a figure which shows an example of a 3rd operation screen. It is a figure which shows an example of a 4th operation screen. It is a figure which shows an example of the 5th operation screen. It is a figure which shows an example of the 6th operation screen. It is a figure which shows an example of the 7th operation screen. It is a figure which shows an example of the 8th operation screen. It is a figure which shows an example of the 9th operation screen. It is a figure where it uses for description about operation at the time of editing a keyword and a category.

  Hereinafter, an information processing apparatus, an information processing method, and an information processing program according to embodiments of the present invention will be described with reference to the drawings. In all the drawings for explaining the embodiments, common constituent elements are denoted by the same reference numerals, and repeated explanation is omitted.

  As illustrated in FIG. 1, the information processing apparatus 1 includes a first instruction unit 11 that instructs an arbitrary explanatory variable, an acquisition unit 12 that acquires an information source corresponding to the arbitrary explanatory variable, and an analysis that analyzes the information source. Unit 13, a first waveform generation unit 14 that generates a first waveform based on a result analyzed by the analysis unit 13, a second instruction unit 15 that instructs an arbitrary objective variable (hereinafter referred to as KPI), A second waveform generation unit 16 that generates a second waveform based on the relevance (correlation, coincidence, etc.) between the KPI waveform and the first waveform, and an output unit 17 that outputs the second waveform to the display unit 2 are provided. .

  The details of an arbitrary explanatory variable instructed by the first instruction unit 11 will be described later. For example, a keyword or a category associated with a plurality of keywords is applicable. The arbitrary explanatory variable may be a phenomenon such as weather information.

  The information source acquired by the acquisition unit 12 is a source that can be accessed via the Internet, and is, for example, a website provided by a newspaper company or a television station, a personal blog, Twitter (registered trademark), or the like. Further, the information source is not limited to the text base. For example, the information source may be an image (moving image and still image), sound, or the like. The acquisition unit 12 digitizes images and sounds and outputs the digitized information to the analysis unit 13. For example, the acquisition unit 12 can extract the transition of the number of appearance posts every day for an image with an arbitrary hash tag.

  The analysis unit 13 has a function as a classifier that performs quantification of information included in the acquired information source. As shown in FIG. 2, the analysis unit 13 includes all or part of a noise removal processing unit 13a, a word count processing unit 13b, and a P / N determination unit 13c.

  As a combination of the analysis unit 13, as shown in FIG. 2A, a form including a noise removal processing unit 13a, a word count processing unit 13b, and a P / N determination unit 13c, and as shown in FIG. In addition, as shown in FIG. 2 (c), there are a mode including a noise removal processing unit 13a and a word count processing 13b and a mode including a noise removal processing unit 13a and a P / N determination unit 13c.

  The noise removal processing unit 13a performs noise removal processing on information (for example, text information) included in the information source. Noise removal processing includes spam classification, advertisement classification, semantic classification, and the like.

  The word count processing unit 13b performs processing for classifying the text data from which noise has been removed by the noise removal processing unit 13a into keyword classification or category classification (divided into sentences).

  The P / N determination unit 13c determines whether the text data from which noise has been removed by the noise removal processing unit 13a or the text data classified by the word count processing unit 13b is positive or negative. . Specifically, it is determined whether the text data after morphological analysis is positive or negative based on machine learning or a large amount of corpus.

  Here, the first waveform will be described with reference to FIG. The first waveform generation unit 14 generates a waveform (first waveform) indicating the transition of time from the result analyzed by the analysis unit 13. For example, as shown in FIG. 3, the first waveform shows the number of occurrences of each word when a certain word (information A1, information A2, information A3 in the example shown in FIG. 3) is designated as an arbitrary explanatory variable. It is a waveform which shows time transition. Moreover, the 1st waveform generation part 14 may produce | generate the thing which put together several words as one 1st waveform instead of the waveform for every word.

For example, if the number of keywords in information A1 is WC _A1 , the number of keywords in information A2 is WC _A2 , the number of keywords in information A3 is WC _A3, and so on, and the number of keywords in information AJ is WC _AJ. The first waveform can be expressed by equation (1).
First waveform = ft (WC _AJ ) (1)

  As will be described later, when the KPI is a stock price of an arbitrary company X, the information A1 to AJ may include a product name related to the company X or a product name not related to the company X. In some cases. Furthermore, the information A1 to AJ may indicate a certain phenomenon (for example, weather information) as well as a word.

Further, the first waveform generation unit 14 may generate the first waveform based on reputation analysis instead of the word count. For example, the weight of the article A included in the information source acquired by the acquisition unit 12 is set as W _a1 , the weight of the article B is set as W _a2 , the weight of the article C is set as W _a3, and the weight of the article i is set as W _ai and then, the weight of the information A1 for any of the companies X and _{W A1,} the weight of the information A2 for any of the companies X and _{W A2,} the weight of the information A3 for any of the companies X and _{W A3,} ··· any of the companies If the weight of the information Aj with respect to X is W _Aj , the first waveform can be expressed by equation (2).
First waveform = ft (W _ai , W _Aj ) (2)

Note that the weight W _a1 of the article A may be a value indicating the reliability of the information source including the article A, the number of links to the information source, or the number of links multiplied by an arbitrary coefficient. It may be a value, the number of followers for the information source, or a value obtained by multiplying the number of followers by an arbitrary coefficient.

  The KPI instructed by the second instruction unit 15 means a quantitative index for measuring the degree of achievement of the target of the organization, business, or business. For example, the stock price, the sales of any product, the number of web accesses, the number of visitors , Audience rating, etc. The second instruction unit 15 receives KPI time-series data and converts it into a KPI waveform. The KPI waveform converted by the second instruction unit 15 is input to the second waveform generation unit 16 and the output unit 17.

  Next, the second waveform will be described with reference to FIG. The second waveform generation unit 16 obtains the relationship between the KPI waveform and each first waveform using multivariate analysis, statistical approach, nonlinear analysis, or processing by financial mathematics (hereinafter referred to as predetermined processing). .

  Note that the second waveform generation unit 16 performs an interpolation process using data that does not include a missing value when any target variable has a missing value. In the interpolation processing, interpolation by copying data not including missing values, interpolation using an average value of a plurality of data not including missing values, or nonlinear interpolation is performed. In addition, interpolation using an average value may be a value obtained by simply averaging a plurality of past data (for example, 10 data), or may be a value calculated by weighting each data (weighted average method). Good. The second waveform generation unit 16 generates a second waveform by substituting the interpolated value for the missing value.

<How to find the degree of agreement>
The second waveform generation unit 16 determines n, which is the number of points constituting the waveform, according to the shape of the KPI waveform, and compares the characteristic of the n point of the KPI waveform with the characteristic of the n point of the first waveform. The second waveform is generated based on the comparison result. That is, the second waveform generation unit 16 dynamically determines n based on the periodicity of the KPI waveform.

  For example, the second waveform generation unit 16 determines n as 7 when the KPI waveform has a periodicity every 7 days. The second waveform generation unit 16 compares the characteristics of the seven consecutive points of the KPI waveform with the characteristics of the seven consecutive points of the first waveform. The characteristic may be a first derivative (slope) or a second derivative (curvature). Further, the second waveform generation unit 16 may generate the second waveform by obtaining the degree of coincidence from the comparison result, may generate the second waveform by obtaining the correlation coefficient, or P / N The determination may be performed to generate the second waveform.

Further, the second waveform generation unit 16 includes the slope of each of the n points (points P ₁₁ to P _1n ) of the KPI waveform and the slope of each of the n points (points P _{21 to} P _2n ) of the first waveform based on the information A1. Are compared with each other to obtain a degree of coincidence, and a plurality of points P _{21 to} P _{2n of} the first waveform determined to have a high degree of coincidence are more than a plurality of points P ₁₁ to P _{1n of the} KPI waveform. If it is ahead in time, it can be determined that the first waveform based on the information A1 is a leading indicator of the KPI waveform. Further, if any of a plurality of points P _{21 to} P _{2n of} the first waveform determined to have a high degree of coincidence is temporally after any of a plurality of points P ₁₁ to P _{1n of the} KPI waveform, the information A1 It can be determined that the first waveform based on is a lag indicator of the KPI waveform.

As illustrated in FIG. 4, the second waveform generation unit 16 generates a second waveform by multiplying the number of keywords of the information A1 by the obtained degree of coincidence. Similarly, as shown in the equation (3), the second waveform generation unit 16 generates the second waveform by multiplying the number of other information keywords (WC) by the obtained degree of coincidence. .
Second waveform = WC × degree of coincidence (3)

  In the example shown in FIG. 4, only two second waveforms are shown. However, as many second waveforms as the number designated by the first instruction unit 11 are generated.

<Other than match level>
In the above description, the method for obtaining the degree of coincidence between the KPI waveform and the first waveform has been described. However, the method is not limited thereto. For example, the correlation coefficient between the KPI waveform and the first waveform may be obtained, and the number of other information keywords (WC) may be multiplied by the obtained correlation coefficient to generate the second waveform. . Also, P / N determination is performed between the KPI waveform and the first waveform, and the number of other information keywords (WC) is multiplied by the obtained P / N determination results to generate the second waveform. May be.

Similarly, the second waveform generation unit 16 also multiplies the degree of coincidence, the correlation coefficient, and the like for the first waveform shown by the expression (2), and changes W _ai and W _Aj every moment. Two waveforms are generated. W _ai that changes momentarily, at any company X the KPI is evaluated by medium 180 for transmitting a variety of information on the Internet (e.g., stock prices) weight sources affecting (weight value) It can be said that there is. Further, W _Aj that changes momentarily, any company X the KPI is evaluated by medium 180 for transmitting a variety of information on the Internet (e.g., stock prices) affects the information (e.g., product) It can be said that it is a weight (weight value).

  The information processing apparatus 1 configured as described above indicates the change in the performance (KPI) by indicating the transition of the second waveform that is considered to have an influence on the KPI waveform, and the reason for the change in the performance. It is possible to show in real time how much it is considered to affect

  In addition, the first instruction unit 11 designates an arbitrary explanatory variable by arbitrarily selecting one or a plurality of explanatory variables from a plurality of predetermined explanatory variables, or inputs an explanatory variable. Thus, a configuration may be used in which an arbitrary explanatory variable is indicated.

  Specifically, one or more keywords or categories can be selected by the keyword / category selection item 120 on the screen shown in FIG.

  Therefore, the information processing apparatus 1 shows the transition of the plurality of second waveforms affecting the KPI waveform, and the reason why the performance (KPI) fluctuates, and how much the reason affects the fluctuation of the performance. Can be shown in real time.

  Moreover, the 1st instruction | indication part 11 may be the structure which instruct | indicates the arbitrary explanatory variables by selecting the said category, where the some explanatory variable is put together in one category.

  A category indicates a concept or rule common to a plurality of explanatory variables. The category may be created by a user freely combining arbitrary explanatory variables, or may be created by combining a plurality of explanatory variables based on a thesaurus (thesaurus).

  For example, even if an individual explanatory variable has a small influence on KPI, a category including the explanatory variable may have a large influence on KPI. In such a case, selecting a category is more efficient than selecting individual explanatory variables.

  In this embodiment, when an explanatory variable has a correlation with KPI and the explanatory variable moves non-negligible with respect to a change in KPI, “the explanatory variable has an influence on KPI. It is expressed as “large”. Therefore, the term “influence” used in the present embodiment does not actually mean that a certain explanatory variable has an influence on the change of KPI.

  Here, the operation of the information processing apparatus 1 will be described. For example, in the software provided by SaaS (Software as a Service), the user performs a predetermined operation (login operation) to display the screen illustrated in FIG. 5 (hereinafter referred to as a post-login screen). .

  The post-login screen is composed of an item 101 for displaying a predetermined logo and project name, and a KPI tab. In the example shown in FIG. 5, a case where two types of KPIs are input is shown. The KPI (sales) tab 102 is a tab in which a KPI of sales of a certain product or store is input. The KPI (membership) tab 103 is a tab in which the KPI of the number of members of a certain company is input. FIG. 5 shows a case where the KPI (sales) tab 102 is selected. Note that the names of tabs such as KPI (sales) and KPI (number of members) can be freely changed by the user.

  A plurality of views are arranged as submenus on the KPI (sales) tab 102. Each view includes a dashboard 104, a KPI (sales) 105, a sentiment 106, a keyword appearance count 107, and a keyword edit 108, but is not limited thereto.

  A movable part (widget) can be selected and displayed on the dashboard 104 as a portal summarizing each view. Each widget is a shortcut to each view of KPI (sales) 105, sentiment 106, and keyword appearance count 107.

<Display period change function>
As illustrated in FIG. 1, the information processing apparatus 1 includes a period changing unit 18 that changes a display period. The period change unit 18 outputs the changed period to the acquisition unit 12. The acquiring unit 12 acquires an information source corresponding to an arbitrary explanatory variable based on the period changed by the period changing unit 18.

  Specifically, the dashboard 104 has a display period changing function 109. The display period changing function 109 can change the display periods of the KPI (sales) 105, the sentiment 106, and the keyword appearance count 107 in a lump by changing the display period. For example, the start date of the display period may be fixed, but the end date may be updated every day to the date of the previous day. In addition, a link for selecting “the last 7 days” and “the last 30 days” may be arranged under the period selection calendar.

  The dashboard 104 displays a logo 110 (hereinafter referred to as a sentimenter 110). The sentimenter 110 is an index indicating whether the KPI can be explained by a plurality of arbitrary explanatory variables instructed by the first instruction unit 11. The shape, color, movement, etc. of the sentimenter 110 change depending on whether or not the explanation can be made. The user can evaluate the relationship (KPI explanation rate) between the KPI and a plurality of arbitrary explanatory variables simply by looking at the sentimenter 110. In addition, the user is expected to select a plurality of arbitrary explanatory variables every day so that the KPI explanation rate becomes high.

<Text display function>
As shown in FIG. 1, the information processing apparatus 1 compares the second waveform generated by the second waveform generation unit 16 with the waveform (KPI waveform) of an arbitrary objective variable instructed by the second instruction unit 15. And an extraction unit 19 that extracts an arbitrary explanatory variable (for example, a keyword) affecting the KPI waveform under a certain condition. The keywords extracted by the extraction unit 19 are input to the output unit 17. As described above, the word “influence” does not mean that a certain keyword actually affects the change of the KPI waveform.

  For example, the extraction unit 19 compares the sentiment score of the previous day with the sentiment score of the current day for all the selected keywords, the keyword with the highest increase rate, the keyword with the lowest increase rate, and the descent The keyword with the highest rate or the keyword with the highest descent rate is extracted. In addition, the extraction unit 19 compares today's sentiment scores for all selected keywords, and extracts the keyword having the highest score or the keyword having the lowest score.

  Specifically, the dashboard 104 has a text display function 111 that displays an arbitrary explanatory variable extracted by the extraction unit 19. The text display function 111 displays an explanatory variable that most affects the KPI compared to the previous day among a plurality of arbitrary explanatory variables. For example, the text display function 111 displays “the keyword that has risen most compared to the previous day is a beer garden”. The displayed explanatory variable may be an explanatory variable directly designated (selected or input) by the user, or may be an explanatory variable automatically selected every day.

Further, the text display function 111 displays all the keywords extracted by the extraction unit 19 (for example, the keyword having the highest rate of increase, the keyword having the lowest rate of increase, the keyword having the highest rate of decrease, The keyword with the highest rate,
The keyword with the highest sentiment score for today and the keyword with the lowest sentiment score for today) may be displayed sequentially, or a different keyword may be displayed each time it is updated.

  The dashboard 104 includes a keyword influence display item 112 that displays an influence of a keyword that is an arbitrary explanatory variable, a KPI summary display item 113 that displays a summary of KPI, a KPI transition display item 114 that displays a transition of KPI, A sentiment display item 115 for displaying a sentiment (second waveform shown in FIG. 4 described above) and a keyword appearance count display item 116 for displaying the number of appearances of a keyword (first waveform shown in FIG. 3 described above) are arranged. Is done. Note that the above only shows an example of items displayed on the dashboard 104. Therefore, only some of the items described above may be arranged on the dashboard 104, or items other than the items described above may be arranged. Hereinafter, the second waveform may be referred to as a sentiment or a sentiment score.

<Influence of keywords>
As shown in FIG. 1, the information processing apparatus 1 selects a plurality of second waveforms from a second waveform generated by the second waveform generation unit 16 under a certain condition, and serves as a basis for each second waveform. In addition, an arbitrary explanatory variable related to the first waveform is obtained, and a determination unit 20 that determines the rank of the arbitrary explanatory variable is provided. The determination unit 20 inputs the determined rank of any explanatory variable to the output unit 17. For example, the determination unit 20 selects ten second waveforms having high sentiment scores, obtains keywords related to the first waveform that is the basis of each second waveform, and determines the ranking of the keywords. In addition, the determination unit 20 selects ten second waveforms having a high sentiment score increase rate, obtains keywords related to the first waveform that is the basis of each second waveform, and determines the ranking of the keywords. Further, the determination unit 20 selects ten second waveforms having a high descending rate of the sentiment score, obtains keywords related to the first waveform that is the basis of each second waveform, and determines the ranking of the keywords.

  Specifically, the keyword influence display item 112 includes a graph 112a that displays keywords (for example, the top 10) having a high sentiment score during a predetermined period, and a keyword ( For example, a graph 112b that displays the top 10) and a graph 112c that displays keywords (for example, the top 10) having a high descending rate of the sentiment score during a predetermined period. The predetermined period is, for example, the most recent month or the most recent week.

  Further, the keyword influence display item 112 may display a message such as “no data to be displayed” without displaying anything in the three graphs at the time of first login. When logging in for the first time, the text display function 111 may display a message such as “Let's input or upload KPI and keywords and start sentiment analysis”.

  Further, the edit widget 112d in the keyword influence display item 112 can be operated to display the edit widget. By using the edit widget, change the display format (for example, change the bar graph to a pie chart), change the display period (for example, change the most recent week to the most recent month), download each graph as an image, and Each graph can be downloaded in CSV format.

  KPI summary display item 113 includes target achievement rate, previous day ratio, previous week (7 days) ratio, previous month (30 days) ratio, KPI explanation rate, target achievement ratio previous month (30 days) ratio, same display period (30 days) The ratio and the display period average are displayed. The KPI summary display item 113 displays “−−−.−−%” or the like when a value cannot be obtained. Note that the order of the items can be changed.

<KPI (sales)>
Next, a display screen when KPI (sales) 105 is selected in the submenu will be described with reference to FIG. The KPI (sales) 105 includes a KPI (sales) display item 117 and a summary table display item 118. The KPI (sales) display item 117 is an item for displaying the KPI instructed by the second instruction unit 15.

  As illustrated in FIG. 1, the information processing apparatus 1 specifies a selected position (for example, date) on the KPI waveform when any of the KPI waveforms displayed on the display unit 2 is selected. 21 and an information generation unit 22 that generates information corresponding to the position (for example, date) specified by the specifying unit 21 based on the KPI instructed by the second instruction unit 15. The information generation unit 22 outputs the generated information to the output unit 17.

  Specifically, as shown in FIG. 6, an example is shown in which the stock price of an arbitrary company X is displayed as the KPI. Here, when somewhere in the KPI waveform is selected with a mouse (if the display unit 2 is a touch panel type, it is selected with a finger), information corresponding to the selected date is read and detailed information 117a is displayed in a pop-up. The For example, the detailed information 117a displays “Project 1 Company X Stock Price (Yen) 2014/6/28 (SAT) 3,695”. When the display period is changed by the display period changing function 109, the horizontal axis of the KPI (sales) display item 117 is changed according to the change.

  An edit widget can be displayed by operating the edit button 117b in the KPI (sales) display item 117. By using the edit widget, change the color of the graph, change the display range of the graph, change the range of the vertical and horizontal axes, change to the logarithmic axis, download the graph as an image, and download the graph in CSV format You can download it at

  Further, the KPI (sales) display item 117 may be configured to display a line graph of the target achievement rate in two axes together with the KPI waveform. In the case of this configuration, a line graph of the target achievement rate may appear as long as the target value is input even for one day during the display period. In addition, when the target value is not input for one day during the display period, the line graph of the target achievement rate may not be displayed. Moreover, even if the logarithmic axis is selected, the biaxial target achievement rate graph does not change. Further, in the editing widget, it may be possible to select non-display of the target achievement rate.

  Also, by operating the KPI (sales) input / edit 117c, the KPI can be input and edited. For example, by operating the KPI (sales) input / edit 117c, the currently displayed KPI can be changed to another KPI.

<Date memo function>
As illustrated in FIG. 1, the information processing apparatus 1 stores an input unit 23 for inputting a message and the message input by the input unit 23 in association with a waveform (KPI waveform) of an arbitrary objective variable for each date. A storage unit 24 and a reading unit 25 that reads a message having a date corresponding to the position specified by the specifying unit 21 from the storage unit 24 are provided. The reading unit 25 outputs the read message to the output unit 17.

  Specifically, a date memo function and a comment function can be added to the KPI (sales) display item 117. The date memo function is a function that can be input and shared for each date from the lower part of the graph of each view, and is common to the time series graphs of each view.

  When no KPI is input, a screen as shown in FIG. 7 is displayed. In this case, it is necessary to input KPI. Therefore, when a KPI file is selected and input, the button 117d is selected to select the file, and the KPI is input. In addition, the KPI is input by directly dropping the file to which the KPI has been input into the area 117e.

  The table 117f shows the items of date and KPI in FIG. 7, but other items may be added. The table 117f shows dates in descending order, and 30 days per page are displayed. In particular, when logging in for the first time, the date may be up to 180 days (6 pages) from yesterday. In addition, when a page transition is attempted without saving, a confirmation pop may be displayed. In addition, after saving, an “analysis update” button (not shown) may be operated. Also,

  Further, the user interface (UI) may be a calendar 117g as shown in FIG. 8 instead of the table 117f. Each item of the calendar 117g may be displayed in the order of date, KPI, target value, and memo, for example.

  When the date is clicked, a balloon 117h is displayed in a pop-up. In the balloon 117h, KPI, target value, and memo can be input. When saving of the balloon 117h is selected, the input content is saved and reflected in the corresponding date item of the calendar 117g. In addition, when there is a date for which KPI is not input, the item of the corresponding date may be highlighted with a highlight or the like. Also, the current day (today) and the previous day (yesterday) in use may be clearly indicated next to the date, and may be visually noticeable by adding a frame color or the like.

<Aggregation table>
The summary table display item 118 is an item for displaying a summary table as shown in FIG. KPI time-series data that is the basis of the KPI waveform is input to the tabulation table. In the summary table, large data and small data can be recognized at a glance. For example, using the heat map function, an item for which data having a large numerical value is input has a darker background color, and an item for which data having a small numerical value is input has a lighter background color. For example, the horizontal axis of the tabulation table is configured in a time series like the sentiment and the number of keyword appearances, and can be expanded or folded by a click operation. As the time series, for example, a cumulative total (a grand total across calendar years), a calendar year, a quarter (1-March, April-June, July-September, October-December), a month, a day, and the like can be selected. The vertical axis of the tabulation table includes, for example, a target value, a KPI (label name), and a target achievement rate.

  Next, a display screen when the sentiment 106 is selected in the submenu will be described with reference to FIGS. 9 to 11. The sentiment 106 includes a KPI (sales) display item 117, a sentiment display item 115, a source list display item 119, and a keyword / category selection item 120. The source list display item 119 includes a tabulation table 119a, an information source 119b, and a keyword related diagram 119c.

<Keyword category>
The keyword / category selection item 120 corresponds to the first instruction unit 11 and is an item for selecting a keyword or a category. The keyword / category selection item 120 can be opened and closed by operating the open / close button 120a. The default is open. When in the closed state, the graph displayed on the KPI (sales) display item 117 or the like can be viewed with a wider width. Further, by selecting the keyword editing button 120b, editing such as addition, change, and deletion of keywords and categories can be performed.

  The keyword corresponds to, for example, a product name, a person name, or a term. For example, 70 keywords are displayed, but are not limited to 70 words. The order in which the keywords are arranged may be sorted, for example, in descending order of influence on the KPI. The keyword may be prepared in advance, or the user may freely input and register the keyword.

<Keyword / category automatic input function>
The information processing apparatus 1 includes a selection unit 26 that selects an explanatory variable by machine learning. The acquisition unit 25 acquires an information source corresponding to an arbitrary explanatory variable instructed by the first instruction unit 11 and an information source corresponding to an arbitrary explanatory variable selected by the selection unit 26. Moreover, the structure which changes an explanatory variable for every predetermined period may be sufficient as the selection part 26. FIG.

  For example, when selecting 70 keywords, the user designates 40 keywords by arbitrarily selecting or inputting keywords, and the remaining 30 are designated by machine learning as keywords that are likely to affect KPI. It may be a simple configuration. In the case of the configuration, the 30 specified by machine learning may be sequentially selected every predetermined period.

  A category indicates a concept or rule common to a plurality of keywords. The category may be created by a user freely combining arbitrary keywords, or may be created by combining a plurality of keywords based on a thesaurus (thesaurus). For example, even if an individual keyword has a small influence on KPI, a category including the keyword may have a large influence on KPI. In such a case, selecting a category is more efficient than selecting individual keywords.

  When a keyword is selected in the keyword / category selection item 120, the second waveform (sentiment) of the keyword is displayed in the sentiment display item 115.

  When “all” is selected in the keyword / category selection item 120, all keywords and categories displayed in the keyword / category selection item 120 are selected, and the second waveform (sentiment) of the selected keyword / category is selected. ) Is displayed in the sentiment display item 115. When no keyword is selected, the second waveform (sentiment) is not displayed in the sentiment display item 115.

  Also, a different color may be specified for each keyword, and the second waveform may be displayed in the same color as the keyword color. According to the said structure, there exists a merit which becomes easy to grasp | ascertain which keyword corresponds to which 2nd waveform at a glance.

  Further, a system color may be determined for each category. For example, “category: brand” is designated by a blue color, and “category: season” is designated by a red color.

<Display method when a category is selected>
When the category is selected, the output unit 17 outputs the second waveform based on all the explanatory variables belonging to the category to the display unit 2, and is a single sum of the second waveforms based on all the explanatory variables belonging to the category. A waveform is output to the display unit 2, a single waveform obtained by averaging the second waveforms based on all explanatory variables belonging to the category is output to the display unit 2, or a second waveform based on all explanatory variables belonging to the category May be calculated by a predetermined calculation method, and one waveform obtained from the calculation result may be output to the display unit 2.

  Specifically, in the keyword / category selection item 120, when a category is selected, all the second waveforms (sentiments) of the keywords belonging to the category may be displayed in the sentiment display item 115. A configuration in which one waveform obtained by summing up the second waveforms (sentiments) of the keywords belonging to the keyword may be displayed in the sentiment display item 115, or one average of the second waveforms (sentiments) of the keywords belonging to the categories may be used. The configuration may be such that the waveform is displayed in the sentiment display item 115, or the second waveform (sentiment) of the keyword belonging to the category is calculated by a predetermined calculation method, and one waveform obtained from the calculation result is the centimeter. May be configured to be displayed in the item display item 115.

<Display of information sources>
Further, in the sentiment display item 115, when an arbitrary place of the second waveform is selected with a mouse (if the display unit 2 is a touch panel type, it is selected with a finger), a balloon 115a pops up as shown in FIG. Is displayed. The balloon 115a displays information (for example, “product name A1 2014/6/28 (sat) adjusted sentiment 70”) that is the basis of the second waveform and the date and time.

  The output unit 17 may be configured to display an information source having a strong influence on the second waveform in association with the second waveform. For example, when the user wants to know the basis for the change in the second waveform, the user can display the information source in association with the second waveform by performing a predetermined operation. Thus, the detailed information of the information source can be obtained instantaneously, so that the user can know the cause of the change in the second waveform.

  Specifically, a plurality of menus are displayed in a balloon 115a so as to be selectable. The menus are, for example, “display only affiliation category”, “display only related keywords”, “display information sources”, “display keyword related diagrams”, and “exclude from graph”.

  The sentiment display item 115 displays only the category to which the keyword of the selected second waveform belongs when “display only belonging category” is selected. For example, the sentiment display item 115 currently includes a second waveform of the product name A1, a second waveform of the product name A2, a second waveform of the product name A5, a second waveform of the product name B1, and a second waveform of the product name B5. When the second waveform of the product name A1 is selected, the second waveform of the product name A1, the second waveform of the product name A2, the second waveform of the product name A2 is the category to which the product name A1 belongs. Only the waveform, the second waveform of product name A3, the second waveform of product name A4, and the second waveform of product name A5 are displayed.

  The sentiment display item 115 displays only the second waveform of the keyword related to the keyword of the selected second waveform when “display related keyword only” is selected.

  The source list display item 119 displays the information source of the selected keyword of the second waveform when “display information source” is selected. The source list display item 119 displays a keyword related diagram when “display keyword related diagram” is selected. The keyword relation diagram is a diagram showing the relevance between keywords in terms of distance or the like.

  The sentiment display item 115 excludes the selected second waveform from the display target when “exclude from graph” is selected.

  Further, the selection location is not limited to one location, and a plurality of selection locations may be selected. When a plurality of locations are selected, only the menu described above may be displayed in the balloon 115a.

  As shown in FIG. 9, the tabulation table 119a is a table showing the result of tabulating the keywords of the second waveform displayed in the sentiment display item 115 on a daily basis. Data for the graph display period can be downloaded in CSV format.

<Source of information>
The information source 119b is an information source obtained by extracting the keyword of the second waveform displayed in the sentiment display item 115 as shown in FIG. The information source includes a text, an acquisition date and time, and a reliability of the information source. Note that the information sources are arranged in descending order of reliability (for example, the number of followers). In addition, the text may be displayed up to a predetermined number of characters (for example, 140 characters), and if the user wants to browse more, the text may be displayed by clicking “More”.

  For example, when the user wants to know the basis of the change of the second waveform displayed in the sentiment display item 115, the user selects the place where the second waveform is changed with a mouse (display unit). When 2 is a touch panel type, it is selected with a finger). When “display information source” of the balloon 115a displayed in the vicinity of the selection is selected, the detailed information of the information source of the keyword of the second waveform is displayed on the information source 119b arranged at the bottom of the same screen. The user can know the basis of the change by browsing the information source 119b.

<Keyword-related figure>
The output unit 17 displays a related diagram generated based on the relevance of an arbitrary explanatory variable on the display unit 2. Specifically, the keyword relation diagram 119c shows a distance between the selected keyword and a keyword related to the keyword.

  For example, in the keyword association diagram 119c, the degree of association and similarity between keywords are calculated by TF (Term Frequency) -IDF (Inverse Document Frequency) or the like, and a keyword association diagram is generated based on the calculation. Also, the keyword relation diagram 119c generates a relation diagram with the selected keyword as the main word. The size of the circle may be calculated based on the appearance frequency. The length of the line may be calculated from the Cos similarity.

  In the above description, the relationship diagram between keywords has been described. However, the present invention is not limited to this, and a relationship diagram between categories may be used, or a relationship diagram between keywords and categories may be used.

<Keyword appearance count>
Next, a display screen when the keyword appearance count 107 is selected in the submenu will be described with reference to FIG. The keyword appearance count 107 includes a keyword appearance count display item 116, a source list display item 119, and a keyword / category selection item 120. The source list display item 119 includes a tabulation table 119a, an information source 119b, and a keyword related diagram 119c.

  The output unit 17 outputs the first waveform generated by the first waveform generation unit 14 to the display unit 2. Specifically, the keyword appearance count display item 116 displays the keyword appearance count (the first waveform shown in FIG. 3 described above).

  Also, a different color may be specified for each keyword, and the first waveform may be displayed in the same color as the keyword color. According to the said structure, there exists a merit which becomes easy to grasp | ascertain which keyword corresponds to which 1st waveform at a glance.

  Further, a system color may be determined for each category. For example, “category: brand” is designated by a blue color, and “category: season” is designated by a red color.

  Further, the edit widget 116b in the keyword appearance count display item 116 can be operated to display the edit widget. By using the edit widget, change the color of the graph (waveform), change the range of the vertical and horizontal axes, change to the logarithmic axis, download each graph as an image, and download each graph in CSV format can do.

  Also, in the keyword appearance count display item 116, when an arbitrary location of the first waveform is selected with a mouse (if the display unit 2 is a touch panel type, it is selected with a finger), a balloon 116b is displayed as shown in FIG. Displayed in a pop-up. The balloon 116b displays information such as the keyword and the number of appearances based on the first waveform (for example, “product name A1 appearance number: 50 2015/5/20 (wed)”).

  The balloon 116b displays a plurality of menus so that they can be selected. The menu includes, for example, “display only affiliation category”, “display only related keywords”, “display information sources”, “display keyword related diagrams”, and “exclude keywords”.

  The keyword appearance count display item 116 displays only the category to which the keyword of the selected first waveform belongs when “display only belonging category” is selected. The keyword appearance count display item 116 displays only the first waveform of a keyword related to the selected keyword of the first waveform when “display related keywords only” is selected.

  The source list display item 119 displays the information source of the selected keyword of the first waveform when “display information source” is selected. The source list display item 119 displays a keyword related diagram when “display keyword related diagram” is selected. The keyword relation diagram is a diagram showing the relevance between keywords in terms of distance or the like.

  The keyword appearance count display item 116 excludes the selected first waveform from the display target when “exclude from graph” is selected.

<Keyword editing>
As illustrated in FIG. 1, the information processing apparatus 1 includes a filter unit 27 that filters arbitrary explanatory variables instructed by the first instruction unit 11.

  The filter unit 27 has a white list in which keywords to be processed are registered and a black list in which keywords that are not to be processed are registered. When a keyword is instructed by the first instruction unit 11, the filter unit 27 refers to the white list or black list and passes the keyword to be processed. The keyword that has passed through the filter unit 27 is input to the acquisition unit 12.

  In addition, the filter unit 27 has a registration list that is not to be processed when the keyword instructed by the first instruction unit 11 is difficult to acquire meaningful information. This registration list may be determined according to requirements such as being less than a predetermined number of characters or a specific character being continuous, or may be extracted from the number of appearances of a keyword and the characteristic value of its transition. The filter unit 27 refers to the registration list when a keyword is instructed by the first instruction unit 11 and satisfies a predetermined requirement such as when the number of characters of the keyword is equal to or less than the predetermined number of characters. Pass through. The keyword that has passed through the filter unit 27 is input to the acquisition unit 12.

  The white list, black list, and registration list described above can be edited by adding, changing, or deleting keywords. The keyword that has passed through the filter unit 27 is registered in the keyword / category selection item 120.

  Here, a display screen when keyword editing 108 is selected in the submenu will be described with reference to FIGS. 13 and 14. In the keyword editing 108, editing such as addition, change, and deletion of keywords and categories can be performed.

  There are multiple ways to edit keywords and categories. For example, the keyword is input by operating the button 108a and selecting a file in which the keyword is input. Further, the keyword is input by directly dropping the file in which the keyword is input into the area 108b. The file format is the CSV format, but is not limited to this. In addition, keywords can be registered up to 70 words, but are not limited to 70 words.

  Further, the keyword and the category may be edited by inputting the category into the input window 108c, entering the keyword into the input window 108d, and selecting the registration button 108e. When the registration button 108e is selected, keyword filtering by the filter unit 27 described above is performed.

  The category and keyword input as described above are reflected in the keyword / category selection item 120 shown in FIG.

  Here, a specific example of editing keywords and categories will be described with reference to FIG. In the following, it is assumed that six keywords have already been registered.

  Enter “brand” in the category and enter “product name A1” in the keyword belonging to the category.

  Further, by selecting “add co-occurrence or co-occurrence exclusion word”, it is possible to specify the co-occurrence keyword and the co-occurrence keyword to be excluded from the input keyword. In the example shown in FIG. 14, “XXX” is designated as the co-occurrence word, and “YYY” is designated as the co-occurrence word to be excluded.

  Further, the keyword editing 108 has a function of automatically saving the input contents after a certain period of time during editing work. In the case of automatic saving, the date and time of automatic saving are displayed.

  Further, when “return to the previous” is selected, for example, the editing state can be selected retroactively to the fifth generation.

  In addition, in this embodiment, the configuration and operation of the information processing apparatus 1 for showing in real time the reason why the performance has changed and how much the reason has an influence on the change in the performance has been described. However, the present invention is not limited to this, and each component may be provided and configured as a method and a program for showing in real time why the performance has changed and how much the reason affects the change in the performance.

  Further, it is realized by recording a program for realizing each function constituting the information processing apparatus 1 on a computer-readable recording medium, causing the computer system to read and execute the program recorded on the recording medium. May be.

  Specifically, the program is an information processing program, and includes a first instruction step for indicating an arbitrary explanatory variable, an acquisition step for acquiring an information source corresponding to the arbitrary explanatory variable, and analyzing the information source. An analysis step, a first waveform generation step for generating a first waveform based on a result analyzed by the analysis step, a second instruction step for indicating an arbitrary objective variable, a waveform of the arbitrary objective variable and the first waveform Is a program for realizing, by a computer, a second waveform generation step for generating a second waveform based on the relevance to and an output step for outputting the second waveform to the display unit 2.

  Furthermore, the “computer system” here includes an OS and hardware such as peripheral devices. The “computer-readable recording medium” refers to a storage device such as a portable medium such as a flexible disk, a magneto-optical disk, a ROM, and a CD-ROM, and a hard disk built in the computer system.

  Furthermore, “computer-readable recording medium” means that a program is dynamically held for a short time, like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. It is also possible to include one that holds a program for a certain time, such as a volatile memory inside a computer system that becomes a server or client in that case. Further, the program may be for realizing a part of the above-described functions, and may be capable of realizing the above-described functions in combination with a program already recorded in the computer system. .

  DESCRIPTION OF SYMBOLS 1 Information processing apparatus, 11 1st instruction | indication part, 12 Acquisition part, 13 Analysis part, 13a Noise removal process part, 13b Word count process part, 13c P / N determination part, 14 1st waveform generation part, 15 2nd instruction part , 16 Second waveform generation unit, 17 output unit, 18 period change unit, 19 extraction unit, 20 determination unit, 21 identification unit, 22 information generation unit, 23 input unit, 24 storage unit, 25 reading unit, 26 selection unit, 27 Filter part, 101 1st item, 102 KPI (sales) tab, 103 KPI (number of members) tab, 104 dashboard, 105 KPI (sales), 106 sentiment, 107 keyword appearance frequency, 108 keyword edit, 108a, 117d Button, 108b, 117e area, 108c, 108d input window, 108e registration button, 109 display period Change function, 110 logo, 111 text display function, 112 keyword influence display item, 112a, 112b, 112c graph, 113 KPI summary display item, 114 KPI transition display item, 115 sentiment display item, 115a, 117h balloon, 116 keyword appearance Number display item, 117 KPI (sales) display item, 117a detailed information, 117b edit button, 117c KPI (sales) input / edit, 117f table, 117g calendar, 118 summary table display item, 119 source list display item, 119a summary table 119b Information source, 119c Keyword related diagram, 120 Keyword / category selection item, 120a Open / close button, 120b Keyword edit button.

Claims (21)

A first instruction unit for indicating an arbitrary explanatory variable;
An acquisition unit for acquiring an information source corresponding to the arbitrary explanatory variable;
An analysis unit for analyzing the information source;
A first waveform generation unit that generates a first waveform based on a result analyzed by the analysis unit;
A second instruction unit for designating an arbitrary objective variable;
A second waveform generating unit that generates a second waveform based on the relationship between the waveform of the arbitrary objective variable and the first waveform;
An information processing apparatus comprising: an output unit that outputs the second waveform to a display unit.   The second waveform generation unit determines n, which is the number of points constituting the waveform, according to the shape of the waveform of the arbitrary objective variable, the characteristics of the n point of the waveform of the arbitrary objective variable, The information processing apparatus according to claim 1, wherein the second waveform is generated based on a comparison result with a characteristic of n points of one waveform. It has a period change section that changes the display period,
The period change unit outputs the changed period to the acquisition unit,
The information processing apparatus according to claim 1, wherein the acquisition unit acquires an information source corresponding to the arbitrary explanatory variable based on the period changed by the period changing unit.   The second waveform generated by the second waveform generation unit is compared with the waveform of the arbitrary objective variable instructed by the second instruction unit, and the waveform of the arbitrary objective variable is affected under a certain condition. The information processing apparatus according to claim 1, further comprising an extraction unit that extracts an arbitrary explanatory variable that gives   A plurality of second waveforms are selected from the second waveforms generated by the second waveform generation unit under certain conditions, and an arbitrary explanatory variable related to the first waveform that is the basis of each second waveform is obtained. The information processing apparatus according to claim 1, further comprising a determination unit that determines a rank of an arbitrary explanatory variable. When a waveform of an arbitrary target variable displayed on the display unit is selected, a specifying unit that specifies a selected position in the waveform of the arbitrary target variable;
The information processing apparatus according to claim 1, further comprising: an information generation unit that generates information corresponding to the position specified by the specifying unit based on an arbitrary objective variable instructed by the second instruction unit. An input section for inputting a message;
A storage unit that stores the message input in the input unit in association with a waveform of an arbitrary objective variable for each date;
The information processing apparatus according to claim 6, further comprising: a reading unit that reads a message having a date corresponding to the position specified by the specifying unit from the storage unit.   The first instructing unit designates the arbitrary explanatory variable by arbitrarily selecting one or a plurality of explanatory variables from a plurality of predetermined explanatory variables, or inputs an explanatory variable. The information processing apparatus according to claim 1, wherein the arbitrary explanatory variable is designated by.   The information processing apparatus according to claim 8, wherein the first instruction unit includes a plurality of explanatory variables grouped into one category, and indicates the arbitrary explanatory variable by selecting the category. The output unit, when a category is selected,
Outputting a second waveform based on all explanatory variables belonging to the category to the display unit;
Outputting one waveform summed up the second waveforms based on all explanatory variables belonging to the category to the display unit;
Outputting one waveform obtained by averaging the second waveforms based on all explanatory variables belonging to the category to the display unit;
Alternatively, the information processing apparatus according to claim 9, wherein the second waveform based on all explanatory variables belonging to the category is calculated by a predetermined calculation method, and one waveform obtained from the calculation result is output to the display unit.   The information processing apparatus according to claim 9, wherein the first instruction unit has an editing function for adding, changing, and deleting arbitrary explanatory variables and categories. A selection unit for selecting explanatory variables by machine learning;
The said acquisition part acquires the information source corresponding to the arbitrary explanatory variables instruct | indicated by the said 1st instruction | indication part, and the information source corresponding to the arbitrary explanatory variables selected by the said selection part. Information processing device.   The information processing apparatus according to claim 12, wherein the selection unit changes the explanatory variable for each predetermined period.   The information processing apparatus according to claim 1, wherein the output unit displays the information source having a strong influence on the second waveform in association with the second waveform.   The information processing apparatus according to claim 13, wherein the information sources are arranged in descending order of reliability.   The information processing apparatus according to claim 1, wherein the output unit displays a related diagram generated based on the relevance of an arbitrary explanatory variable on the display unit.   The information processing apparatus according to claim 1, wherein the output unit outputs the first waveform generated by the first waveform generation unit to the display unit.   The information processing apparatus according to claim 1, wherein the second waveform generation unit performs an interpolation process using data that does not include a missing value when an arbitrary objective variable has a missing value.   The information processing apparatus according to claim 1, further comprising a filter unit that filters any explanatory variable instructed by the first instruction unit. A first indicating step for indicating an arbitrary explanatory variable;
An acquisition step of acquiring an information source corresponding to the arbitrary explanatory variable;
An analysis step of analyzing the information source;
A first waveform generation step of generating a first waveform based on the result analyzed by the analysis step;
A second indicating step for indicating an arbitrary objective variable;
A second waveform generation step of generating a second waveform based on the relationship between the waveform of the arbitrary objective variable and the first waveform;
An information processing method comprising: an output step of outputting the second waveform to a display unit. A first indicating step for indicating an arbitrary explanatory variable;
An acquisition step of acquiring an information source corresponding to the arbitrary explanatory variable;
An analysis step of analyzing the information source;
A first waveform generation step of generating a first waveform based on the result analyzed by the analysis step;
A second indicating step for indicating an arbitrary objective variable;
A second waveform generation step of generating a second waveform based on the relationship between the waveform of the arbitrary objective variable and the first waveform;
An information processing program for realizing the output step of outputting the second waveform to the display unit by a computer.

Images