KR20150055132A - System for inference of emotional language using ahp and fuzzy integral - Google Patents

Abstract

The present invention relates to an emotional language inference system using AHP and fuzzy integral and, more specifically to an emotional language inference system capable of analyzing, by using AHP, specific situations of positive and negative dispositions in emotionally expressed language which is classified by a psychological emotion model, and inferring emotional language applied with a users′ subjective dispositions by applying the analyzed specific situations to a fuzzy integral. The present invention collects contexts which can be obtained from a user′s mobile device calculates an importance by analyzing the collected contexts with the AHP; calculates a fuzzy measure according to the calculated importance; and infers emotional language of the mobile user by applying the calculated fuzzy measure and emotionally expressed language to the fuzzy integral.

Description

FIELD OF THE INVENTION [0001] The present invention relates to a system and method for estimating emotional language using AHP and fuzzy integral,

The present invention relates to a system capable of inferring emotional language using AHP and fuzzy integral.

More specifically, the tendency toward the positive / negative aspects of the emotional expression language classified in the psychological emotional model is analyzed by analyzing the specific situation using the Analytic Hierarchy Process (AHP), and the fuzzy integral (Fuzzy Integral) To a system capable of deducing an emotional language that reflects the user's subjective tendency.

The research for expressing emotion is largely divided into the study of psychological emotion model and the emotion engineering research model.

The Wheel of emotion model proposed by Plutchik as a psychological emotional model consists of 8 basic emotions and basic emotions of Joy, Trust, Fear, Surprise, Sadness, Disgust, Anger and Anticipation, . Ekman 's Ekman Basic Emotion model classifies emotions that are commonly applied regardless of culture and race, and suggests a model consisting of six emotions that basically feel in similar situations. Robinson categorized emotion into six categories and classified them into positive / negative aspects. Parrot classifies emotions using three levels of tree structure, and subdivides the six emotions of Love, Joy, Surprise, Anger, Sadness, and Fear.

However, the classification of research and emotion is mainly based on context - aware reasoning technique which is needed for personalization.

Recently, research on personalization service through awareness of emotional information such as emotion has been actively carried out in order to provide personalization service considering user 's emotion through studying personalization service through context recognition in mobile environment.

As an example of the personalization service field technique, a method and system for inferring optimal service in a ubiquitous environment using context information and emotion recognition are disclosed in Japanese Patent Application Laid-Open No. 10-2010-0100380.

And more particularly, to a service inference method and system using context information and emotion recognition that can identify a user's state and emotion such as a user's action or position and provide a service required by a user in consideration of the status and feelings of the user.

In this case, the user's emotion information is collected from the user's facial expressions and voice information, and the user's situation information is collected from the user's identification number, location, and motion information, and then, based on the probability-based reasoning using the Bayesian network, Recognize and fuse the situation, and provide services to the user based on the result.

The above-described technique collects information from the ubiquitous environment, so that a ubiquitous environment must be established. In a place where such an environment is not established, situation information can not be recognized and emotion recognition can not be deduced.

As another example, in Patent Publication No. 10-0822029, a personalization service method and system using user history in a mobile terminal are described.

The present invention provides a personalization service method and system that collects interaction and situation information generated between a user and a mobile terminal and manages it as a user history, and has a user differentiation based on the user history.

At this time, event information is collected from the phone number, music, photograph, TV, e-mail, and Internet access information of the mobile terminal, and the position, state, emotion, and event occurrence time of the user are detected when the event information is generated. Then, the emotion state of the user is inferred from the conversation voice pattern of the user or the pulse of the bio-signal, and the user's behavior is inferred from the vibration of the mobile terminal.

In order to provide a personalized service in the mobile terminal, the above-described technology collects and infer the status and emotion information of the user and displays representative images corresponding to the collected status information. However, there is a problem that the emotion information of the user can not be collected.

That is, since the emotion information of the user differs depending on the user, depending on the time, place, and situation, it is difficult to collect the emotion information unless the user inputs the emotion information.

Also, the above technique can not infer common emotion information that is leveled for each user. For example, it is possible to infer emotional information only when the emotional information accumulates (learns) to some extent by collecting event information such as position and time with respect to one user, and it takes a long time for learning, There is a problem that an error occurs.

Published Patent Publication No. 10-2010-0100380 (2010.09.15.) Patent Registration No. 10-0822029 (Apr. 07, 2008)

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above problems, and it is an object of the present invention to provide an AHP capable of quantitatively analyzing context information obtained from a mobile, And an emotional language reasoning system using a fuzzy integral.

Another problem to be solved by the present invention is to provide an AHP and a fuzzy integral which can acquire emotional information obtainable through a terminal without going through a learning process and deduce an emotional language suitable for providing a personalized emotional service from acquired emotional information To provide an emotional language reasoning system using the sentence.

According to an aspect of the present invention, there is provided a method of analyzing a context that can be acquired from a user mobile, analyzing the collected context using an Analytic Hierarchy Process (AHP), calculating a significance level, And the emotion language of the mobile user is inferred by applying the calculated fuzzy measure and the emotional expression language to a fuzzy integral. The emotional language reasoning system using the AHP and the fuzzy integral, So as to solve the technical problem.

The present invention is applied to a system capable of providing a variety of personalized services according to an inferred emotional language by quantitatively analyzing contextual information obtainable from a mobile and inferring a personalized emotional language based on context information There is an effect that can be.

Further, the inference of the emotional language through the context information according to the present invention is constructed based on the leveled situation information, and can be applied to the user immediately without accumulating information according to learning, and it is possible to derive various emotional languages have.

FIG. 1 is a diagram for explaining the Choke integration for the set X. FIG.
FIG. 2 is a diagram schematically illustrating a main configuration of an emotional language reasoning system using an AHP and a fuzzy integral according to an embodiment of the present invention. Referring to FIG.
3 is a diagram illustrating a process for inferring emotional language in an emotional language reasoning system using AHP and fuzzy integral according to an embodiment of the present invention.
4 is an AHP hierarchical diagram of the emotional expression language classified by EARL in the emotional language reasoning system using AHP and fuzzy integral according to the present invention.
FIG. 5 is a graph showing frequency of selection of positive emotion expression in environmental and behavioral aspects.
FIG. 6 is a diagram showing weights of the relative importance of the emotional expression language viewed from the positive side.

The terms and words used in the present specification and claims should not be construed as limited to ordinary or dictionary terms and the inventor may properly define the concept of the term in order to best describe its invention It should be construed as meaning and concept consistent with the technical idea of the present invention.

Therefore, the embodiments described in the present specification and the configurations shown in the drawings are merely the most preferred embodiments of the present invention and are not intended to represent all of the technical ideas of the present invention. Therefore, various equivalents And variations are possible.

Before describing the present invention with reference to the accompanying drawings, it should be noted that the present invention is not described or specifically described with respect to a known configuration that can be easily added by a person skilled in the art, Let the sound be revealed.

The present invention relates to a system capable of inferring emotional language using AHP and fuzzy integrals. The present invention relates to a system for inferring emotional language by using AHP (Analytic Hierarchy Process) The present invention relates to a system for analyzing a specific situation and applying a fuzzy integral to a specific situation analyzed to infer an emotional language reflecting a user's subjective inclination.

Before describing the contents of the present invention in detail, the technology used in the present invention will be described first, and then the constitution of the present invention using this technology will be described.

AHP will be described.

The analytic hierarchy process (AHP) is a multi-criteria decision making model developed by Saaty, which is a systematic analysis of the decision-making process. This is a model for deriving a rational result by performing pairwise comparison and performing pairwise comparisons on the overall evaluation items and calculating relative importance to the evaluation items.

In particular, AHP supports quantitative analysis of qualitative problems, supports decision-making systematically, organizes complex and unclear problems into multiple hierarchies, and partial relationships are analyzed through pairwise comparisons By evaluating the significance of the value of the product on a quantified scale, allowing decision makers to make more accurate decisions.

At this time, it can be classified into 5-point scale and 9-point scale according to the numerical scale. In the case of 9-point scale, it can be defined as shown in Table 1 below.

importance Justice One Similar 3 Slightly important 5 Important 7 very important 9 Extremely important 2, 4, 6, 8 Median of defined values

At this time, when the pair comparison matrix is A and the element (evaluation item) of A is a _ij , a _ij can be estimated using the following equation (1).

Where n is the number of elements in the hierarchy to be compared and w _i is the relative importance of the n elements.

Therefore, matrix A with an element a _ij can be expressed as:

Here, A represents a pair of comparison matrices, and a _ij represents an element of the pair comparison matrix A.

In addition, the weight vector w = [w ₁ , w ₂ , w ₃ , ... , w _n ] can be expressed as A · W = n · W.

Where n is the eigenvalue of the matrix A and W is the eigenvector of A.

On the other hand, if the respondent does not consistently respond to the relative importance of each item in the pairwise comparison matrix, the accuracy of the pairwise comparison matrix becomes low. Therefore, the consistency index (CI) and the consistency ratio (CR) should be used to verify whether the importance of pairwise comparison is consistent.

The consistency index (CI) and the consistency ratio (CR) can be obtained by the following Equations (2) and (3).

Here, CI is the coherence index, [lambda] _max is the maximum eigenvalue of the pair comparison matrix, and n is the number of elements to be compared in one layer.

Here, CR is the coherence rate, CI is the coherence index, and RI is the random index (RI).

At this time, the coherence index (CI) has a smaller value as the response of the respondent becomes coherent. If the ratio is less than 0.1, the coherence index is determined to be consistent.

That is, in the present invention, the contexts that can be acquired from the mobile device are collected, and the importance of each context is analyzed using the AHP, so that the relative importance of the context can be calculated.

Fuzzy theory, on the other hand, is a mathematical approach to the process of decision making by the brain in various problems in ambiguous and unclear situations.

It is based on the idea of fuzzy sets introduced by Professor L.A. Zadeh of the University of California at Berkeley in 1965.

The concept of a fuzzy set can be mathematically expressed from binary logic that each object belongs to or does not belong to a certain group by representing the degree to which each object belongs to the group as a membership function.

The fuzzy measure is a mathematical representation of the relationship between a and A, by indicating the ambiguity of the fact that the ambiguous element a in the general set A belongs to the subset P of A. Briefly summarized, it can be said that the number of cases is taken into account and the movement is designated accordingly.

In the present invention, the ambiguity of the emotional language of the user's subjective judgment using the interaction coefficient fuzzy measure (fuzzy measure) and the fuzzy integral (fuzzy integral) to analyze and classify the relationship between the specific situation and the emotional expression language And quantitatively processing the emotional language value and deducing it.

The fuzzy integral is explained in more detail.

In order to apply the fuzzy integral, a fuzzy measure must be calculated. In the present invention, an interaction coefficient fuzzy measure (? Fuzzy Measure) using an interaction coefficient may be applied.

The interaction coefficient fuzzy measure is widely applied as an evaluation model of the subjective measure by the fuzzy measure proposed by SUGENO, and is generally defined by the following equation (4).

Here, A and B are sets, and g _λ is a fuzzy measure with an interaction coefficient applied.

The interaction coefficient fuzzy measure has the same characteristics as the following equation (5) according to the value of [lambda] (interaction coefficient), and is applied as a model of the subjective measure.

Here, A, B are set, _λ g is a fuzzy measure, g _λ (A) the interaction factor is applied fuzzy measure, g _λ (B) of the A set is a measure of the fuzzy set B.

The fuzzy integral is an integral application method for the fuzzy measure. The fuzzy integral applied in the present invention quantitatively handles the ambiguity of the subjective judgment of the empirical expression language and the evaluator by using the Choquet Integral, which is one of the fuzzy integrals Lt; / RTI >

The Choquet Integral is a Functional defined by the French mathematical Choquet on the Capacity and is defined as in Equation (6) below.

Where f is a real number function on X and g is a fuzzy measure on X.

Hereinafter, the Choquet Integral will be described in detail.

If a simple function f on X is given by the following Equation 7 with respect to the set X = {x ₁ , x ₂ , x ₃ , x ₄ }

로 나타낸다.However, 0 < r ₁ < r ₂ <r ₃ <r ₄ , D _i ∩D _j = ø (i ≠ j)

Respectively.

Here, i and j are integers.

This is illustrated in FIG.

FIG. 1 is a diagram for explaining the Choke integration for the set X, and the Choquet Integral for the fuzzy measure g of the single function f can be expressed by the following equation (8).

로 표현된다.only,

Lt; / RTI &gt;

Where A is a set, i is an integer, and g (A) is a fuzzy measure for set A.

That is, as shown in the figure, the Choke integral means integrating along the vertical axis representing the real value f that satisfies the additive.

The emotional language reasoning system using the AHP and the fuzzy integral according to the present invention collects contexts obtainable from the user mobile, analyzes the collected contexts by AHP, calculates importance, calculates a fuzzy measure according to the calculated importance, The empirical language of the mobile user is inferred by applying the fuzzy measure and emotional expression language to the fuzzy integral.

2 is a block diagram schematically showing a main configuration of an emotional language reasoning system 20 using an AHP and a fuzzy integral according to an embodiment of the present invention. The context collecting unit 21, the AHP analyzing unit 22, An emotion expression language storage unit 24, a fuzzy integration unit 25, and an emotional language detection unit 26. The emotion expression language storage unit 24,

3 is a diagram illustrating a process for inferring emotional language in an emotional language reasoning system using AHP and fuzzy integral according to an embodiment of the present invention.

Hereinafter, the present invention will be described with reference to Figs. 2 and 3.

The emotional language reasoning system 20 using the AHP and the fuzzy integral according to the present invention may be constituted by an application installed in the mobile 10 or a server connected to the mobile 10 in a connected manner.

The context collecting unit 21 collects a context obtainable from the mobile 10.

Here, the mobile device 10 refers to a mobile device such as a smart phone and a tablet PC, which can be carried by a user and moved with the user.

The context obtainable from such mobile 10 is limited.

For example, the context obtainable from the mobile 10 can be limited to information such as location from the GPS provided in the mobile 10, behavior of the mobile user, application installed in the mobile, weather and time.

At this time, the context detection according to the behavior of the mobile user can be inferred from several examples. For example, the user's behavior can be judged as driving when the user is able to drive and the mobile position is moved by a considerable distance, and if the travel distance is short, it can be judged by jogging or walking (walking). Also, if a specific application is running through an installed application, it can be judged that the mobile is operating, and if the location of the mobile is collected as a camping site, it can be determined that the user's behavior is camping.

That is, the behavior of the user can be configured to be inferred based on the location of the mobile, the type of the installed application, etc. based on the sex, age, hobby, specialty and various qualification information of the user.

In the context, the weather can be collected as weather information of the location based on the location on the collected mobile 10, and time can be collected from the mobile 10 or standard time information.

The context that can be collected from the mobile 10 can be added as the development of the sensor, the emotional modeling develops. For example, if it is configured to be able to collect heart rate, respiration, body temperature, etc. from the user in connection with the user's body, more accurate user's sensibility can be recognized.

As described above, the context collected from the mobile 10 can be utilized as a kind of emotion that is empirically perceived in a specific situation, and the relationship between the collected contexts can be expressed differently for each user and is various and complex.

In order to quantify the collected contexts as universal data, we surveyed each context and analyzed it through AHP.

The AHP analyzer 22 compares the collected contexts with each other for each context to calculate a weight for importance, and Table 2 is a result of calculating the weights of the collected contexts.

location Mobile Apps User behavior weather time weight location One 7 3 6 8 0.513 Mobile Apps One 1/3 5 6 0.137 User behavior One 7 7 0.262 weather One 3 0.054 time One 0.032

Looking at [Table 2], it was found that the location is seven times more important than mobile apps (mobile apps), three times more than user behavior, six times more than weather and eight times more important than time.

In addition, mobile apps were found to be one - third times more important than user behavior, five times more than weather and eight times more important than time, and user behavior was seven times more important than weather and time.

Therefore, APH analysis of location, mobile app, user behavior, weather, and time resulted in weights of 0.513, 0.137, 0.262, 0.054, and 0.032, respectively.

The weight of context through the survey was analyzed as having importance in order of location, user behavior, mobile app, weather and time.

The fuzzy measure calculator 23 is for calculating the fuzzy measure and is calculated using the importance (AHP weight) calculated by the AHP analyzer 22 and the interaction coefficient (?).

In this case, the fuzzy measure can be calculated by using the calculated importance and the interaction coefficient, and the singleton fuzzy measure ratio can be calculated using the Singleton Fuzzy Measure Ratio Standard.

The interaction coefficient (?) Means a correlation between contexts. In the present invention, a value of 3.0001 is used.

Table 3 below shows the fuzzy measure calculated by the importance and interaction coefficient calculated through AHP.

Sets Fuzzy Measure (Fuzzy Measure) {} 0.0 {Location} 0.301327 {Mobile App} 0.080934 {Location, Mobile App} 0.455425 {User Action} 0.153889 {Location, User Action} 0.594330 {Mobile App, User Action} 0.272189 {Location, Mobile App, User Action} 0.819571 {Weather} 0.031895 {Location, Weather} 0.362054 {Mobile App, Weather} 0.120574 {Location, Mobile App, Weather} 0.530898 {User Action, Weather} 0.200509 {Location, User Action, Weather} 0.683093 {Mobile App, User Action, Weather} 0.330129 {Location, Mobile App, User Action, Weather} 0.929887 {Time} 0.018501 {Location, Time} 0.336553 {Mobile App, Time} 0.103928 {Location, Mobile App, Time} 0.499204 {User Action, Time} 0.180932 {Location, User Action, Time} 0.645818 {Mobile App, User Action, Time} 0.305798 {Location, Mobile App, User Action, Time} 0.883561 {Weather, Time} 0.052166 {Location, Weather, Time} 0.400651 {Mobile App, Weather, Time} 0.145767 {Location, Mobile App, Weather, Time} 0.578866 {User Action, Weather, Time} 0.230140 {Location, User Action, Weather, Time} 0.739509 {Mobile App, User Action, Weather, Time} 0.366953 {Location, Mobile App, User Action, Weather, Time} 1.0

Thirty-two combinations can be made according to location, mobile application, user behavior, weather and time, and Table 3 shows the fuzzy measure calculated using singleton fuzzy measure non-standard by the combination of these contexts.

According to [Table 3], the combination of two contexts yields the highest fuzzy measure in the combination of location and user behavior, and when combining three contexts, the highest fuzzy measure is the location, mobile application and The highest number of combinations of user behavior was calculated.

The emotion expression language storage unit 24 stores and manages a language that can represent the emotion of the user and can be configured as a classified emotion expression language.

Expression language is defined and classified by several people. Among them, the classification model used in the present invention can use the EARL proposed by HUMAINE to express emotional expression language as an XML-based ontology.

The EARL is a model that classifies 48 emotional expression languages into 10 categories including positive / negative tendencies, and schematically shown in FIG. 4.

3 is a diagram illustrating an AHP hierarchical level of an emotional expression language in an emotional language reasoning system using an AHP and a fuzzy integral according to the present invention.

Here, a survey was conducted to calculate the relative importance of emotional expression language.

FIG. 5 is a graph showing frequency of selection of positive emotion expression in environmental and behavioral aspects.

In the graph of FIG. 5, the X-axis represents the frequency of the emotional expression language selected by the questioner, and the Y-axis represents the group name of the emotional expression language classified by the EARL.

In FIG. 5, the emotion expression language corresponding to the Caring category in FIG. 4 is "Affection", "Empathy", "Friendliness" and "Love".

Also, in the case of emotional expression language corresponding to the positive and lively category, the frequency of "Delight", "Excitement", "Happiness", "Joy" and "Pleasure"

In addition, "Relaxed" and "Relieved" were selected as the emotional expression languages related to the place in the case of the Quiet positive category. Especially, it was confirmed that most positive emotional expression language was selected in most perceived situations.

The emotional expression language of EARL uses 48 languages in 10 categories, but the emotional expression language used in the present invention is not derived from the questionnaire and excludes the emotional expression language having a low frequency of use.

A total of 37 emotional expression languages were extracted, and AHP was performed to calculate the importance for each of 37 emotional expression languages.

FIG. 6 is a diagram showing weights of the relative importance of the emotional expression language viewed from the positive side.

According to FIG. 6, the five most emotional expression languages that are most positively evaluated are Love (L: 0.22), Trust (L: 0.082), Happiness (L: 0.074) , Friendliness (L: 0.070).

Hurt, Doubt, Fear, Sadness, Anger, etc., which have a lower weight of synthetic weight, correspond to emotional expression language of negative aspect as the weight is lower.

Thus, the size of the positive / negative tendency toward emotional expression language can be expressed relatively by using AHP.

In order to apply the AHP according to the design conditions, the relative importance of each emotional expression language may be configured to be arbitrarily set by the user, and when the user is configured to set it arbitrarily, the value of the synthetic weight value may be changed.

The fuzzy integral unit 25 performs a fuzzy integral by applying a fuzzy integral to the calculated fuzzy measure and the emotional expression language. At this time, the fuzzy integral can be performed using the Choquet Integral.

The results of one embodiment calculated according to the fuzzy integral are shown in Table 4 below.

Mo.
Emotional language
Location Mobile App User Action Weather Time Fuzzy integral
(Fuzzy Integral) One Pleasure 60 70 40 15 10 44.0566 2 Boredom 50 50 60 10 20 44.9616 3 Happiness 60 40 45 15 10 42.6303 4 Delight 50 30 50 30 20 41.1855 5 Irritation 45 40 40 25 5 37.3979 6 Amusement 50 50 35 5 5 36.4185 7 Joy 40 60 40 15 5 36.4068 8 Excitement 40 40 35 10 5 32.4158 9 Relaxed 50 20 30 20 20 31.9698 10 Calm 45 25 30 20 20 31.5894 11 Satisfaction 35 40 35 15 10 31.4455 12 Fear 45 10 35 20 35 30.0956 13 Love 30 45 35 5 15 28.2995 14 Interest 30 35 30 15 5 26.9971 15 Affection 40 20 30 5 5 26.2501 16 Content 20 40 40 30 5 24.9715 17 Annoyance 30 25 20 10 5 21.6289 18 Worry 35 15 20 5 15 21.3272 19 Sadness 30 20 20 10 5 20.8584 20 Stress 25 15 35 5 5 20.6779 21 Friendliness 20 35 30 5 5 20.4201 22 Tension 35 10 20 5 10 19.8810 23 Embarrassment 35 10 15 5 5 18.0960 24 Politeness 30 10 20 5 5 18.0544 25 Surprised 25 15 20 5 5 17.6740 26 Anger 15 20 30 5 5 16.0955 27 Disgust 15 10 15 5 5 12.0695 28 Trust 10 20 20 5 5 11.8197 29 Relived 20 15 5 10 5 11.4383 30 Hope 20 10 5 5 5 10.2904 31 Disappointment 10 10 15 5 5 9.86730 32 Hurt 5 15 10 5 5 6.76562 33 Pride 5 5 10 5 5 5.76945 34 Shame 5 5 10 5 5 5.76945 35 Shock 5 10 5 5 5 5.40467 36 Doubt 5 10 5 5 5 5.40467 37 Envy 5 5 5 5 5 5.00000

As can be seen from the results in [Table 4], as the value of the integrated evaluation value is larger, it can be seen as an emotional language for the collected context. In other words, the collected context is collected based on context recognition, and can be viewed as an emotional language for the current situation of the user.

The emotional language detection unit 26 detects the emotional language according to the user's context (collected context) from the fuzzy integral value performed by the fuzzy integration unit 25, and performs a function of detecting a large value among the values of the integrated evaluation value do.

It is a matter of course that the fuzzy integration unit 25 and the emotion language detection unit 26 can be configured in a unified structure according to the design conditions.

Table 5 below shows 11 items with a large integrated value of the fuzzy integral in the result of [Table 3] and the group classified by the ELRL to which each item belongs, and the fuzzy language detection unit 26 The emotional language having the greatest integral value may be inferred as the emotional language of the user and the sensed emotional language may be displayed in the emotional language display unit 30. [

No. Emotional language context  Relative importance Groups categorized by EARL One Pleasure 0.042 positive and lively 2 Boredom 0.013 negative and passive 3 Happiness 0.074 positive and lively 4 Delight 0.048 positive and lively 5 Irritation 0.005 negative and forceful 6 Amusement 0.021 positive and lively 7 Joy 0.016 positive and lively 8 Excitement 0.040 positive and lively 9 Relaxed 0.038 quiet positive 10 Calm 0.012 quiet positive 11 Satisfaction 0.020 positive thoughts

[Table 4] Referring to Table 4, the sensible language inferred based on the user's current situation information is 'pleasure' and the group to which 'pleasure' belongs belongs to 'Positive and lively' among the 10 groups classified by EARL .

As described above, the emotion-based language reasoning system using the AHP and the fuzzy integral according to the present invention enables the emotion-based language reasoning according to the user's situation using the context collected from the mobile.

In other words, context and emotional language are calculated as relative importance reflecting user's subjective tendency through AHP, and it is calculated as fuzzy integral through fuzzy measure, so that inference of emotional language suitable for user's situation It is possible.

In addition, weights of AHP and weights of emotional expression language are quantified as standardized data through questionnaires, so that it is possible to quantitatively evaluate emotional language deduced, and it is possible to infer emotional language even under difficult conditions of acquisition of emotional language .

Therefore, the present invention can be applied to various fields using the emotional language reasoning system using the AHP and the fuzzy integral.

For example, it can be applied to a system for music recommendation, food recommendation, movie recommendation, and travel destination recommendation corresponding to emotional language according to the inference of emotional language according to the user's situation. Further, the inferred emotional language can be applied to express the face of the robot to express emotions according to the wise situation of the user.

At this time, in order to apply the inferred emotional language to other systems and provide personalized service, it is necessary to distinguish the application target and design should be made considering the age, sex, service target and service environment of the target person. That is, applicable information or classification criteria for the application target should be differentiated.

The present invention is applied to a system capable of providing a variety of personalized services according to an inferred emotional language by quantitatively analyzing contextual information obtainable from a mobile and inferring a personalized emotional language based on context information .

In addition, the inferred emotional language is constructed based on the leveled situation information, so it can be directly applied to the user without accumulating information according to learning, and various emotional languages can be derived.

1 to 6 describe only the main points of the present invention. As far as various designs can be made within the technical scope thereof, the present invention is not limited to the configurations and functions of Figs. 1 to 6 It is self-evident that it is not limited.

10: Mobile
20: Emotional Language Inference System
21: Context collecting unit
22: AHP analysis section
23: Purge measure calculating section
24: Emotion expression language storage unit
25: Fuzzy integral part
26: Emotion language detector
30: emotional language display

Claims (5)

The collected context is analyzed by the Analytic Hierarchy Process (AHP) to calculate the importance, the fuzzy measure is calculated according to the calculated importance, and the calculated fuzzy measure measure and emotional expression language is applied to a fuzzy integral to infer the emotional language of the mobile user. The emotional language reasoning system using AHP and fuzzy integral.
The method according to claim 1,
The context comprises:
The location of the mobile, the behavior of the mobile user, the application installed on the mobile, the weather, and the time, and the AHP and the fuzzy integral based on the fuzzy integral.
The method according to claim 1,
The fuzzy measure is a measure of
AHP (Analytic Hierarchy Process), and the interaction coefficient between the collected contexts is used to calculate the importance of the emotional language inference system using the AHP and the fuzzy integral.
The method according to claim 1,
The emotional expression language includes:
The emotional language inference system using the AHP and the fuzzy integral, which is selected from the emotional expression languages classified into 10 categories by the EARL.
The method according to claim 1,
The fuzzy integral is a function of
A sensory language reasoning system using AHP and fuzzy integral, characterized by being a Choquet Integral.

Images