CN113283943A - Commodity demand prediction-based paper product packaging production management system and method - Google Patents

Abstract

The application discloses a paper product packaging production management system and method based on commodity demand prediction, an action information acquisition module acquires action information of a user on an e-commerce platform, a re-purchased commodity prediction module determines re-purchased commodities of the user according to purchase information in the action information and predicts current re-purchased demand degree of the re-purchased commodities so as to predict current re-purchased commodities of the user, a co-purchased commodity prediction module determines an action segment where an action of the user for purchasing the current re-purchased commodities in the current future from the action information, predicts the co-purchased commodities and the demand degree of the current re-purchased commodities according to the action information in the action segment, and a paper product packaging determination module determines paper product packages of corresponding specifications and the quantity of the paper product packages based on the current re-purchased commodities and the co-purchased commodities. The system can shorten the supply time of paper product packages, and can reduce the paper product inventory and maintenance cost due to targeted pre-production.

Description

Commodity demand prediction-based paper product packaging production management system and method

Technical Field

The application relates to the technical field of paper product production, in particular to a paper product packaging production management system and method based on commodity demand prediction.

Background

The paper product processing industry is one of the important industries in China, both corrugated boards for packaging boxes and paper towels serving as sanitary articles are indispensable daily articles in daily life, and most of packages of products, goods and other articles are wrapped and contained by the packaging boxes formed by paper products.

The logistics warehouse of the express delivery company is used as a delivery site, paper packaging is required to be carried out on commodities before delivery, and then delivery is carried out, wherein the paper packaging is ordered from a packaging box manufacturer. At present, for manufacturers of paper product packages, a large number of paper product packages with different specifications are usually produced in advance and put into inventory so as to supply the paper product packages in time when each delivery site orders, but this can cause occupation of a large amount of inventory space of a carton warehouse, and cartons in inventory can not be ordered by the delivery sites due to less purchase quantity of commodities with corresponding specifications, and then are stacked in the warehouse for a long time and cannot be used, so that the turnover capacity of the cartons in the warehouse is reduced, and the maintenance cost of the cartons is also improved. Therefore, how to reasonably arrange the production of paper packages so that the produced paper packages can at least partially meet the paper package requirements of the delivery site is an urgent problem to be solved at present.

Disclosure of Invention

Based on this, in order to rationally arrange the production of paper package for the paper package of output can at least partially accord with the paper package demand of delivery website, the following technical scheme is disclosed in this application.

In one aspect, a paper product packaging production management system based on commodity demand prediction is provided, including:

the action information acquisition module is used for acquiring action information of a user on the E-commerce platform;

the repurchase commodity prediction module is used for determining the repurchase commodities of the user according to the purchase information in the action information, predicting the current repurchase demand degree of the repurchase commodities and further predicting the current repurchase commodities of the user;

the co-purchased commodity prediction module is used for determining an action segment of an action of the user for purchasing the current re-purchased commodity in the current date from the action information and predicting the co-purchased commodity and the demand degree of the current re-purchased commodity according to the action information in the action segment;

the paper product package determining module is used for determining paper product packages and the number of the paper product packages in corresponding specifications based on the current purchased commodities and the purchased commodities;

and the paper product package production device is used for carrying out corresponding production based on the determined specification parameters and the number of the paper product packages.

In one possible embodiment, the repurchase merchandise prediction module comprises:

the purchasing information extraction unit is used for extracting all purchasing information about the commodities available for repeated purchasing from the action information of the user on the E-commerce platform;

the time interval estimation unit is used for determining the same commodity which is purchased for multiple times and the corresponding purchase time thereof from the purchase information and estimating the purchase time interval of the commodity for the user according to the purchase time;

and the repurchase demand judging unit is used for judging the repurchase demand degree of the user for the commodity according to the latest time of purchasing the commodity by the user and the purchasing time interval.

In a possible implementation manner, the time interval estimation unit is further configured to determine, from the purchase information, similar commodities purchased multiple times and corresponding purchase times thereof, and estimate a purchase time interval of the commodity for the user according to the purchase times;

the repurchase demand judging unit is also used for judging the repurchase demand degree of the user for the commodities according to the latest time when the user purchases the commodities and the purchase time interval.

In one possible embodiment, the co-purchased goods prediction module comprises:

a time interval acquisition unit for respectively acquiring occurrence time intervals of adjacent action occurrence times;

an interval node acquisition unit, configured to perform density-based clustering on the occurrence time intervals, calculate an average value of the occurrence time intervals in each class, and use the occurrence time intervals included in the class with the largest average value as interval nodes;

and the action fragment dividing unit is used for dividing the action occurrence time according to the interval node to obtain action fragments.

In one possible embodiment, the co-purchased goods prediction module further comprises:

the commodity sequence forming unit is used for extracting corresponding commodities in each action fragment to form a plurality of commodity sequences corresponding to the user;

a mapping relation generation unit for generating a mapping relation between commodities in the commodity sequence in a latest set time period;

and the co-purchase commodity prediction unit is used for inputting the mapping relation and the corresponding action into a pre-constructed commodity co-purchase prediction model and predicting the co-purchase commodities and the co-purchase demand degree.

In one possible embodiment, the paper product package determination module comprises:

the demand degree sequencing unit is used for sequencing the current purchased commodities and the purchased commodities together according to respective demand degrees to obtain a demand degree sequence;

the commodity determining unit is used for extracting a plurality of commodities with the highest demand degree from the demand degree sequence, so that paper product packages required by the commodities are just matched with the production capacity of the paper product package production device;

and the package determining unit is used for taking the paper product packages with the corresponding specifications of the plurality of commodities and the number thereof as the determined paper product packages and the number thereof.

On the other hand, the paper product packaging production management method based on commodity demand prediction is further provided, and comprises the following steps:

acquiring action information of a user on an e-commerce platform;

determining the re-purchased commodities of the user according to the purchase information in the action information, predicting the current re-purchased demand degree of the re-purchased commodities, and further predicting the current re-purchased commodities of the user;

determining an action fragment where the action of the user for purchasing the current purchased goods at the current date is located from the action information, and predicting the purchased goods and the demand degree of the current purchased goods according to the action information in the action fragment;

determining paper product packages and the number of the paper product packages of corresponding specifications based on the current purchased commodities and the purchased commodities;

and performing corresponding production based on the determined specification parameters and the number of the paper product packages.

In a possible implementation manner, the determining a repurchase product of the user according to the purchase information included in the action information and predicting the current repurchase demand degree of the repurchase product includes:

extracting all purchase information about the commodities available for repurchase from the action information of the user on the E-commerce platform;

determining the same commodity which is purchased for multiple times and the corresponding purchasing time thereof from the purchasing information, and estimating the purchasing time interval of the user for the commodity according to the purchasing time;

and judging the re-purchasing demand degree of the user for the commodity according to the latest purchasing time of the user for the commodity and the purchasing time interval.

In a possible implementation manner, the determining, according to the purchase information included in the action information, a re-purchased product of the user and predicting a current re-purchase demand of the re-purchased product further includes:

determining the similar commodities which are purchased for multiple times and the corresponding purchasing time thereof from the purchasing information, and estimating the purchasing time interval of the commodity of the type of the user according to the purchasing time;

and judging the re-purchasing demand degree of the user for the commodities according to the latest time of purchasing the commodities by the user and the purchasing time interval.

In a possible implementation manner, the determining, from the action information, an action segment of an action of the user to buy the current purchased goods at the current time includes:

respectively acquiring occurrence time intervals of adjacent action occurrence times;

clustering the occurrence time intervals based on density, calculating the mean value of the occurrence time intervals in each class, and taking the occurrence time intervals contained in the class with the maximum mean value as interval nodes;

and dividing the action occurrence time according to the interval nodes to obtain action fragments.

In a possible implementation manner, the predicting the co-purchased product and the demand thereof of the current re-purchased product according to the action information in the action fragment includes:

extracting corresponding commodities in each action fragment to form a plurality of commodity sequences corresponding to the user;

generating a mapping relation between commodities in the commodity sequence in a latest set time period;

and inputting the mapping relation and the corresponding action into a pre-constructed commodity co-purchasing prediction model to predict co-purchasing commodities and co-purchasing demand degrees thereof.

In a possible embodiment, the determining of the paper product packages and the number thereof with corresponding specifications based on the current purchased products and the purchased products with the same specifications includes:

the current purchased commodities and the purchased commodities with the current purchased commodities are sequenced together according to respective demand degrees to obtain a demand degree sequence;

extracting a plurality of commodities with highest demand from the demand sequence, so that paper packages required by the commodities are just matched with the production capacity of a paper package production device;

and taking the paper product packages with the corresponding specifications of the plurality of commodities and the quantity thereof as the determined paper product packages and the quantity thereof.

In the embodiment, by analyzing the behavior of the user on the e-commerce platform, whether the user purchases the goods repeatedly in a short period or not is predicted according to the goods repeatedly purchased by the user and the purchase time of the goods, and under the condition that the user purchases the goods repeatedly in a short period, the co-purchased goods related to the purchased goods are also predicted, whether the user has the re-purchase demand and the co-purchase demand is judged, and then the required paper package is determined, and the paper package is produced in advance, so that when the user actually purchases the goods after a short period of time and the delivery site orders the corresponding paper package to the manufacturer according to the goods purchased by the user, the manufacturer directly provides the paper package produced in advance according to the prediction result to the express delivery site, the supply time of the paper package can be shortened, the delivery efficiency of the express delivery site is improved, and the paper inventory and maintenance cost are reduced due to the targeted pre-production, so that the stock of the paper product package is more reasonable.

Drawings

The embodiments described below with reference to the drawings are exemplary and intended to be used for explaining and illustrating the present application and should not be construed as limiting the scope of the present application.

Fig. 1 is a block diagram of a structure of an embodiment of a paper product packaging production management system based on commodity demand prediction disclosed in the present application.

FIG. 2 is a demand graph.

Fig. 3 is a schematic flow chart of an embodiment of a paper product packaging production management method based on commodity demand prediction disclosed in the present application.

Detailed Description

In order to make the implementation objects, technical solutions and advantages of the present application clearer, the technical solutions in the embodiments of the present application will be described in more detail below with reference to the drawings in the embodiments of the present application.

An embodiment of the paper product packaging production management system based on commodity demand prediction disclosed in the present application is described in detail below with reference to fig. 1 and 2. As shown in fig. 1, the system disclosed in this embodiment mainly includes: the system comprises an action information acquisition module, a repeated purchase commodity prediction module, a same purchase commodity prediction module, a paper package determination module and a paper package production device.

The action information acquisition module is used for acquiring action information of a user on the E-commerce platform.

The e-commerce platform refers to a platform with an online shopping function, such as a skatecat, a naughty, a kyoto and the like, and the e-commerce platform acquired by the action information acquisition module may be multiple or one of the multiple.

The action information refers to information that the action information acquisition module performs actions related to commodity purchase on any one of the e-commerce platforms, and the purchase of commodities on the e-commerce platform also belongs to one of the actions. The action information R can be expressed as { user U, commodity name G, commodity category C, action O, occurrence time T, e-commerce platform P }, where user U can be a user name; the commodity category C is a category to which the commodity belongs, for example, if the commodity name G is "patent examination manual" 2010, then C is a book; the action O mainly comprises four actions of browsing introduction, paying attention to collection, adding a shopping cart and purchasing; the occurrence time T is the time when the action O starts to occur; the e-commerce platform P is the platform where the action takes place.

If there are multiple e-commerce platforms for acquiring information by the action information acquisition module, the user may select commodities at the multiple e-commerce platforms simultaneously and finally select one of the platforms for purchase, so that the e-commerce platforms are all related to the purchase of the commodities, and when the user performs actions related to the purchase of the commodities (including browsing, paying attention to collection, adding a shopping cart and placing a purchase order) on any one of the e-commerce platforms, corresponding action information is recorded and is uniformly sorted according to the occurrence time of the actions.

Table 1 below is a part of the action information table obtained by the action information obtaining module from the action information of the user U1.

Table 1 partial action information table of user U1 purchasing goods

User' s

Name of commodity

Name of trade company

Categories of goods

Movement of

Time of occurrence

E-commerce platform

U1

srkf1

kfd1

Instant coffee

Browse introduction

Day T-114

P1

U1

srkf1

kfd1

Instant coffee

Add shopping cart

Day T-114

P1

U1

srkf1

kfd2

Instant coffee

Browse introduction

Day T-114

P2

U1

srkf1

kfd3

Instant coffee

Browse introduction

Day T-114

P2

U1

srkf1

kfd1

Instant coffee

Purchasing

Day T-114

P1

U1

srkf1

kfd1

Instant coffee

Purchasing

Day T-78

P1

U1

kfbl1

kfd1

Coffee mate

Browse introduction

Day T-78

P1

U1

kfbl1

kfd2

Coffee mate

Browse introduction

Day T-78

P2

U1

kfbl1

kfd1

Coffee mate

Collection of interest

Day T-78

P1

U1

tch1

kfd1

Soup spoon

Browse introduction

Day T-78

P1

U1

tch1

kfd1

Soup spoon

Collection of interest

Day T-78

P1

U1

srkf1

kfd1

Instant coffee

Purchasing

Day T-47

P1

U1

srkf1

kfd1

Instant coffee

Purchasing

Day T

P1

The table records the introduction of the user to browse commodities, the addition of commodities to a shopping cart, the purchase of commodities and other actions according to the time of starting, and shows that the user U1 browses and purchases instant coffee srkf1 in different merchants of different E-commerce platforms, and finally selects one merchant of one platform to purchase the instant coffee, and purchases the instant coffee repeatedly after 36 days, 67 days and 114 days, and browses coffee partners at the same time after 36 days, but does not purchase but only pay attention to collection. The action information in table 1 is listed in order of the occurrence time of the action, wherein the earlier the occurrence time, the earlier the order. It will be appreciated that only a partial record of the user U1 is intercepted in table 1, and therefore the action information may also include a record of the merchandise information that occurs earlier.

The repurchase commodity prediction module is used for determining the repurchase commodities of the user according to the purchase information in the action information, predicting the current repurchase demand degree of the repurchase commodities and further predicting the current repurchase commodities of the user.

The repurchase product refers to a product which is purchased by a user before and is purchased again for repeated needs, for example, the user has a habit of drinking a certain brand of instant coffee at ordinary times, and the brand of instant coffee is purchased again every time the user drinks the product, so that the brand of instant coffee belongs to the repurchase product.

The current repurchase demand degree refers to the repurchase expectation of the user to the repurchase commodities at the current time, and the magnitude of the repurchase demand degree can be represented by the numerical value within the interval of [0, 1 ]. For example, the time interval between the instant coffee purchase by the user is usually about 30 days, and the user has recently purchased instant coffee, the current time is not yet the time interval of 30 days, so the re-purchase demand degree at the current time is zero; if the current time is close to 30 days from the latest instant coffee purchase, the repeated purchase demand degree at the moment is greater than zero and is at a higher numerical value, and the numerical value is increased along with the increase of the number of days until the user places an order to purchase the instant coffee, and the demand degree value is zero again; if the time interval is exceeded for 30 days and the longer time is exceeded, the user is considered to have abandoned the repurchase of the commodity and the demand count is likewise restored to zero.

The current repurchase commodity is the repurchase commodity with the current repurchase demand degree reaching a certain numerical value, if the demand degree is lower, the user cannot repurchase the commodity in a short time, so the current repurchase commodity is not the repurchase commodity but the expected repurchase commodity, and if the current repurchase commodity is overdue for a long time, the user abandons the repurchase commodity, so the current repurchase commodity is not the repurchase commodity but the abandoned commodity.

The number of the current purchased goods may be one or more, depending on the kind of goods that have been purchased periodically in the action information of the current date of the user.

And the co-purchased commodity prediction module is used for determining an action fragment where the action of the user for purchasing the current re-purchased commodity in the current period is located from the action information, and predicting the co-purchased commodity and the demand degree of the current re-purchased commodity according to the action information in the action fragment.

Because some commodities may have a certain correlation, a user may purchase a commodity related to the commodity after purchasing the commodity based on the correlation, for example, after purchasing a book, if the book is still in a forward or a continuation, the purchase of the forward or the continuation may occur; or auxiliary materials such as nutrient solution, rooting powder and the like are needed to be bought after the water culture plant is bought; or auxiliary seasonings such as coffee mate and the like are found to be needed to be bought after the instant coffee is bought, and the like; in addition, the purchase sequence of some commodities can be changed, such as tea leaves and vacuum cups, the fact that tea leaves are bought to make tea by using the vacuum cups is possible, the fact that tea leaves can be just bought to soak and drink the tea leaves can be found after the vacuum cups are bought, the purchase sequence of the tea leaves and the purchase sequence of the vacuum cups can be different, the purchase correlation exists between the tea leaves and the vacuum cups, namely, after one of the tea leaves is bought, a user may want to buy the other tea leaves, and the commodity purchasing prediction module further predicts the needed commodities of the user through the purchase correlation of the commodities. Like the repurchase of merchandise, the same merchandise will also have its same desirability.

The action segment refers to an action information sequence containing a plurality of continuous action information, and the sequence also contains the current purchasing action for purchasing the current purchased goods. For example, the currently purchased product is instant coffee, and the operation information of all five operations occurring on the day T-114 in table 1 includes the purchasing operation of instant coffee, so that the sequence formed by the five operations is one operation segment Sec1, while the operation information of all six operations occurring on the day T-78 similarly forms another operation segment Sec2, and the purchasing operations occurring on the day T-47 and the day T respectively form operation segments Sec3 and Sec 4.

After the purchased goods prediction module predicts the current purchased goods, the purchased goods prediction module predicts whether the user will purchase other goods having purchase correlation with the current purchased goods when purchasing the current purchased goods according to the action information contained in the action segments occurring in the past period, and if the user is predicted to purchase the goods having purchase correlation, the goods are the purchased goods of the current purchased goods, namely the goods purchased together with the current purchased goods.

And the paper product package determining module is used for determining paper product packages and the number of the paper product packages in corresponding specifications based on the current purchased commodities and the purchased commodities.

Continuing with table 1 as an example, assuming that the current time is T +25 days, and the purchased product prediction module predicts that the user U1 will purchase the current purchased product, that is, a box of instant coffee, in a short period of time, and the purchased product prediction module also predicts that the user will purchase a box of coffee mate at the same time, therefore, for the merchant of the user U1, a paper package capable of separately packaging a box of instant coffee and a box of coffee mate is required, and the paper package determination module determines the specification parameters such as size, shape, ridge shape, layer number and the like based on the specification parameters such as size, shape, characteristics and the like of the product. And for other users, whether the current purchased commodities and the purchased commodities exist again or not can be predicted in the same way, and corresponding paper packages are determined, so that the paper packages required by all users in the near term are obtained.

For example, the user totals Un, and is expected to purchase a book N1 order, a nutrient N2 order, an instant coffee N3 order, and so on. A single item herein refers to a unit purchase amount when the platform is sold, because a single item on the platform may be sold as a single item, or multiple items may be sold as a single item, for example, ten items in one package are sold in one package, and a unit purchase amount is in one package. The corresponding paper packages include a paper package requiring an XH1 model for packaging a single book, a paper package requiring an XH2 model for packaging a single nutrient solution, and the like.

It can be understood that the paper package determining module may classify all the commodities on the e-commerce platform in advance, and screen out commodities that need to be paper packaged before shipment, such as imported commodities and others, where the originally packaged commodities may be shipped directly without repackaging, and some commodities need to be repackaged, such as books, audio and video, potted flowers, food, suitcases, shampoo and others, that need to be repackaged for shipping or unpacked for retail, and then determine whether the currently purchased commodities and the purchased commodities of the user are commodities that need to be repackaged, if so, determine paper packages of corresponding specifications, otherwise, ignore the commodity requirements.

The paper product package production device is used for carrying out corresponding production based on the determined specification parameters and the number of the paper product packages.

After production, paper packages of all users participating in commodity demand prediction are obtained, when the users purchase corresponding commodities according to prediction in a short period, a corresponding paper package is ordered from a paper package producer by a delivery merchant or an express delivery site to package and deliver the commodities, and at the moment, the corresponding paper package is produced in advance by the producer due to the fact that prediction is carried out in advance and according to the prediction result, so that the corresponding paper package can be directly sent to the express delivery site.

In addition, because a manufacturer may have some stocks of paper packages, after the paper package determining module determines the specifications and the quantity of the paper packages, the stock of the paper packages with corresponding specifications can be judged first, and the stock can be counted in the quantity of the paper packages to be produced, so that the paper package producing device only needs to produce the rest parts which cannot be met by the stock and does not need to produce all the paper packages. For example, the package of XH1 paper needs N1 pieces, and the stock quantity is N2 pieces, only (N1-N2) pieces need to be produced.

It should be noted that, the system is directed to a user which is a plurality of users in a set space range, and a plurality of delivery sites and at least one paper product packaging manufacturer exist in the space range, and after determining the paper product packaging quantity required by all users in the space range in the near term, no matter which merchant of which e-commerce platform the user purchases the paper product packaging required in the near term, the paper product packaging required in the near term is not changed by the delivery site selected by the merchant, so for the manufacturer, the difference is only that the delivery site to which the paper product packaging produced in advance is sent is different, for example, instant coffee purchased by the user is delivered by the merchant 1 through the delivery site 1, the merchant site 1 orders corresponding cartons to the manufacturer, and if the merchant 2 delivers the paper product through the delivery site 2, the delivery site 2 orders corresponding cartons to the manufacturer, therefore, only the delivery site adopted by the merchant ordered by the user is located in the range of the supply area of the manufacturer, and the commodity of the merchant is usually not far away from the delivery site, so that the commodity warehouse of the e-commerce platform is located in the range of the supply area.

In the embodiment, by analyzing the behavior of the user on the e-commerce platform, whether the user purchases the goods repeatedly in a short period or not is predicted according to the goods repeatedly purchased by the user and the purchase time of the goods, and under the condition that the user purchases the goods repeatedly in a short period, the co-purchased goods related to the purchased goods are also predicted, whether the user has the re-purchase demand and the co-purchase demand is judged, and then the required paper package is determined, and the paper package is produced in advance, so that when the user actually purchases the goods after a short period of time and the delivery site orders the corresponding paper package to the manufacturer according to the goods purchased by the user, the manufacturer directly provides the paper package produced in advance according to the prediction result to the express delivery site, the supply time of the paper package can be shortened, the delivery efficiency of the express delivery site is improved, and the paper inventory and maintenance cost are reduced due to the targeted pre-production, so that the stock of the paper product package is more reasonable.

In one embodiment, the repurchase product forecasting module comprises a purchase information extracting unit, a time interval estimating unit and a repurchase demand judging unit.

The purchase information extraction unit is used for extracting all purchase information about the repurchase commodities from the action information of the user on the E-commerce platform.

Taking table 1 as the operation information of the user on the e-commerce platform as an example, all the operations are extracted as the operation information of "purchase", and it is determined whether or not all the four pieces of operation information are related to the purchasable item, and as is clear from table 1, all the four pieces of operation information are extracted because all the four pieces of operation information are related to srkf1 instant coffee and belong to the purchasable item. The repurchase product refers to a product which has a periodic repurchase attribute because of consumption when the product is used after purchase, such as some foods and daily necessities, while a product which is collected instead of a daily consumable, such as a gold watch and a commemorative silver coin, does not belong to the repurchase product, and a wearing article such as clothes and shoes does not belong to the repurchase product because of high purchasing randomness of a user.

The time interval estimation unit is used for determining the same commodity which is purchased for a plurality of times and the corresponding purchase time thereof from the purchase information, and estimating the purchase time interval of the commodity for the user according to the purchase time.

The same product refers to a same-name product of the same brand model specification, and can be distinguished as the same product by a product barcode, for example, in this case, "nesto coffee 1+2 extra-strong 390g instant coffee" and "nesto coffee 1+2 milk-flavored 390g instant coffee" are two products other than the same product. The four pieces of purchase information extracted from table 1 are identical in product name and therefore are identical.

The purchase time interval is the time interval of the user for repurchasing the commodity, and if the purchase record occurring on the T-114 day in the table 1 is that the user U1 purchases the srkf1 instant coffee for the first time, three purchase time intervals are provided in the case of repurchasing three times later, and the purchase time intervals are 36 (114-78), 31 (78-47) and 47 according to the sequence of occurrence. The purchase time interval includes a minimum time interval and a maximum time interval, wherein 31 is the minimum time interval for the user U1 to srkf1 instant coffee, and 47 is the maximum time interval for the user U1 to srkf1 instant coffee.

The repurchase demand judging unit is used for judging the repurchase demand degree of the user for the commodity according to the latest time of purchasing the commodity and the purchasing time interval.

The user U1 purchases instant coffee for the last time by T days, if the current time is T +10 days, the current time is 10 days after the user purchases instant coffee for the last time, the current minimum time interval is 31 days away, and therefore the repeated purchase demand degree is zero; if the current time is T +100 days, the current time exceeds 1.2 times of the current maximum time interval (1.2 is a set coefficient), so that the user is judged to abandon the srkf1 instant coffee, and the repurchase demand degree is zero; if the current time is in the interval of [ T +31, T +47], the repurchase demand is greater than zero, and a specific value thereof may be calculated by a curve as shown in fig. 2 or a similar curve disclosure, the function value of the curve starts with T + Tmin as an initial point, the demand gradually increases already at the initial point is greater than zero, finally gradually decreases to T + Tmax, and reaches a maximum value of the demand at a midpoint (Tmax-Tmin)/2, it is understood that since the user may purchase earlier than the current minimum time interval or later than the current maximum time interval, an extension period is provided for a period of time before T + Tmin and after T + Tmax, respectively Tmin-Ta and Tmax-Tb, the extension period may be set to (Tmax-Tmin)/4, where 4 is a predetermined coefficient, for example, (47-31)/4=4 days, then Ta is 27 and Tb is 51.

The time interval estimation unit is further used for determining the similar commodities purchased for multiple times and corresponding purchase time thereof from the purchase information, and estimating the purchase time interval of the commodity of the type of the user according to the purchase time. The repurchase demand judging unit is also used for judging the repurchase demand degree of the user for the commodities according to the latest time when the user purchases the commodities and the purchase time interval.

Assuming that the commercial product purchased on day T in table 1 is srkf2 instant coffee instead of srkf1 instant coffee, and the commercial products purchased in the other three times are not changed, it is still the frequency of purchasing instant coffee for the user U1, but if the time interval and the purchasing demand are calculated according to the same commercial product, the prediction result of purchasing again is wrong, for example, instant coffee is already abandoned when predicted according to the same commercial product. Therefore, the time interval estimation unit calculates the difference between the maximum and minimum time intervals when estimating the time interval for the same product, and calculates the difference for the same product if the difference reaches a certain level (for example, when Tmax >1.5 × Tmin).

The same type of product refers to the same type of product with different names, for example, the aforementioned "Nestle coffee 1+2 extra strong 390g instant coffee" and "Nestle coffee 1+2 milk flavor 390g instant coffee" are the same type of product but not the same product. Assuming that srkf1 is strong, srkf2 is milk-flavored, and the current time is T +39 days, the user's current repurchase demand for instant coffee is at a maximum of 1.

In one embodiment, the co-purchased goods prediction module comprises: the device comprises a time interval acquisition unit, an interval node acquisition unit and an action fragment dividing unit.

The time interval acquisition unit is used for respectively acquiring the occurrence time intervals of the adjacent action occurrence times.

For example, in table 1, the occurrence time may be specified as hours and minutes among six actions occurring on day T-78, the six actions in table 1 are also listed in order of occurrence from top to bottom, the six actions occur at five time intervals, and the favorite collection action occurring on day T-78 and the purchase action occurring on day T-47 are also adjacent actions, so there is also an occurrence time interval between the two actions. Generally speaking, all actions have occurrence time intervals, and the intervals are long or short.

The interval node acquisition unit is used for clustering the occurrence time intervals based on density, calculating the mean value of the occurrence time intervals in each class, and taking the occurrence time intervals contained in the class with the maximum mean value as interval nodes.

The time span of the user actually using the e-commerce platform may be long, for example, 10 years of time exist, so that the occurrence time intervals obtained after actual sequencing are many, wherein the occurrence time intervals are necessarily many, the time intervals are the time that the user never browses and purchases the goods on the e-commerce platform, for example, the user browses the goods on the platform every day after 7 days, and browses the goods on the platform every other day after 3 days, the user may not log in any e-commerce platform for browsing and purchasing for many consecutive days, but may log in and browse some goods every hour, and the like, so that the occurrence time intervals are large or small, and the distribution is random. Therefore, all the action occurrence times need to be classified, and the longer interval times and the shorter interval times are distinguished, so that the action information of the user for intensively online shopping within a certain period of time is classified into the same sequence, and the re-browsing actions after the occurrence time interval of a large period are classified into another sequence, so that the commodity correlation can be conveniently obtained by utilizing the online shopping centralization of the user.

Therefore, the occurrence time interval can be clustered and divided through a DBSCAN (sensitivity-Based Spatial Clustering of Applications with Noise) Clustering algorithm or other Density-Based Clustering algorithms, and areas with high enough Density can be divided into clusters, wherein the clusters comprise clusters formed by short interval time generated when goods are intensively browsed and clusters formed by long-term non-landing platforms. By calculating the average value of each cluster class, it is possible to determine which cluster class is the cluster class corresponding to the longest interval time, and then each occurrence time interval included in the cluster class is taken as an interval node.

Taking table 1 as an example, after clustering, a plurality of cluster clusters are obtained, wherein one cluster class is {47d, 31d, 36d }, the average value of the cluster class is the largest, and the average value exceeds 30 days, so each time interval in the cluster class is taken as an interval node.

And the action fragment dividing unit is used for dividing the action occurrence time according to the interval node to obtain action fragments.

And dividing all action information in the table 1 into four action segments by taking the class cluster as an interval node, wherein five actions occurring on the T-114 day are divided into the same action segment, six actions occurring on the T-78 day are divided into another action segment, and the actions occurring on the T-47 day and the T day are divided into separate action segments respectively to obtain four action segments.

In one embodiment, the co-purchased goods prediction module further comprises: the commodity sequence forming unit, the mapping relation generating unit and the co-purchased commodity predicting unit.

The commodity sequence forming unit is used for extracting corresponding commodities in each action fragment to form a plurality of commodity sequences corresponding to the user.

After the action segments are obtained, commodities contained in the action information are extracted and form a commodity sequence, a single user can correspond to a plurality of commodity sequences, one action segment can form one commodity sequence, and partial sequences in the sequences can be concentrated on a certain number of days or uniformly distributed in each time period, which depends on the operation habits of the users and the result of cluster division. Taking table 1 as an example, four commodity sequences are formed.

The mapping relation generating unit is used for generating the mapping relation among the commodities in the commodity sequence in the latest set time period.

The mapping relation can indicate that two commodities have purchase correlation, and two parties in the mapping relation have the characteristic of synchronous browsing, and even synchronous purchase can occur. However, the purchase correlation is also limited in some cases, and if the mapping relationship occurs a long time ago, the purchase action is not generated due to the mapping relationship at the current time, and therefore, the product sequence of the latest set time period of the user is subjected to the mapping relationship extraction, for example, the latest set time period is within the latest 90 days.

Specifically, the mapping relationship generation unit combines adjacent and repeated commodities in the commodity sequence in the latest set time period to obtain a simplified sequence. Taking table 1 as an example, three product sequences exist within the last 90 days, a sequence 1 is formed by six actions of days T-78, a sequence 2 is formed by one action of days T-47, and a sequence 3 is formed by one action of days T. Where only sequence 1 contains more than one action, the adjacent and repeated items in sequence 1 are combined to obtain a simplified sequence of { instant coffee, coffee mate, spoon }.

The mapping generation unit then establishes a mapping between the purchased good and other non-purchased goods in the reduced sequence, including { instant coffee-coffee mate } and { instant coffee-tablespoon }.

And the co-purchased commodity prediction unit is used for inputting the mapping relation and the corresponding action into a pre-constructed commodity co-purchased prediction model and predicting the co-purchased commodities and co-purchased demand degrees thereof.

The commodity co-purchasing prediction model is obtained by pre-training the existing action information of a large number of users, and the co-purchasing demand degree of the users on the co-purchasing commodity coffee mate at the current time is predicted by inputting the mapping relation of the instant coffee and the coffee mate and the corresponding actions (including browsing and concerned collection) of the coffee mate in the sequence 1 into the model. The purchasing demand degree also represents the magnitude of the repurchase demand degree through the numerical value within the interval of [0, 1 ]. Specifically, the farther the occurrence time of the coffee mate in the sequence 1 is from the current time, the smaller the demand degree is, the more the times of browsing and paying attention to the collection are, the larger the demand degree is, and in the process of calculating the demand degree, the weight of paying attention to the collection is greater than the weight of browsing, the demand degree of the coffee mate is finally calculated, and whether the coffee mate is a commodity purchased together is judged according to the demand degree. The mapping relation of { instant coffee-tablespoon } can be input into the commodity purchasing prediction model in a similar way to calculate the tablespoon demand degree.

In one embodiment, the paper product package determination module comprises: the system comprises a demand degree sequencing unit, a commodity determining unit and a packaging determining unit.

And the demand degree sequencing unit is used for sequencing the current purchased commodities and the purchased commodities together according to respective demand degrees to obtain a demand degree sequence.

Because the daily production capacity of the paper product packaging production device is limited, and the demand degree of the purchased commodities and the purchased commodities is more than zero for a plurality of consecutive days, the production time of the paper product packaging needs to be determined. Therefore, under the condition that the demand degrees of the currently purchased commodities and the purchased commodities are more than zero, the commodities with higher demand degrees can be preferentially produced, and the higher the demand degree is, the closer the date of ordering the commodity is to the user is, the more the ordering demand of the delivery site can be received immediately after production is finished, so that the carton demand of the delivery site and the delivery efficiency satisfaction degree of the user can be quickly met.

Specifically, for all the determined products purchased again and the products purchased together, the demand degrees are mixed and sorted from large to small to obtain a demand degree sequence, the demand degrees in the sequence are large or small, and some products cannot be sorted to the front position because the demand degrees are low in the past, so that the corresponding cartons cannot be produced, and the demand degrees gradually become large along with the increase of time, and are also sorted to the front position, so that the corresponding cartons are preferentially determined and produced.

The commodity determining unit is used for extracting a plurality of commodities with the highest demand degree from the demand degree sequence, so that paper product packages required by the commodities are just matched with the production capacity of the paper product package production device.

Assuming that the production capacity of the paper product packaging production device is N, determining the commodity with the highest current demand degree in the sequence according to the sequence from large to small of the demand degree, determining the paper product package of the commodity, further obtaining the productivity required by the paper product package, judging whether the productivity exceeds the production capacity of the production device, extracting the commodity and adding the paper product package of the commodity into a production plan if the productivity does not exceed the production capacity, then determining the commodity with the highest current demand degree from the sequence until the sum of the paper product package corresponding to the commodity with the highest current demand degree and the productivity demands of the paper product packages corresponding to all the commodities extracted before just exceeds the production capacity of the production device, and indicating that all the commodities extracted at present just meet the production capacity.

And the package determining unit is used for taking the paper product packages with the corresponding specifications of the plurality of commodities and the quantity thereof as the determined paper product packages and the quantity thereof.

The package determining unit takes the number of paper product packaging machines corresponding to all the commodities extracted currently as the determined paper product packages and the number of the paper product packages, and then the paper product packaging production device performs typesetting and production on the paper product packages and stores produced cartons.

An embodiment of a method for managing paper packaging production based on commodity demand prediction disclosed in the present application is described in detail below with reference to fig. 3. The present embodiment is a method for implementing the aforementioned embodiment of the paper product packaging production management system.

As shown in fig. 3, the method disclosed in this embodiment includes the following steps:

acquiring action information of a user on an e-commerce platform;

determining the re-purchased commodities of the user according to the purchase information in the action information, predicting the current re-purchased demand degree of the re-purchased commodities, and further predicting the current re-purchased commodities of the user;

determining an action fragment where the action of the user for purchasing the current purchased goods at the current date is located from the action information, and predicting the purchased goods and the demand degree of the current purchased goods according to the action information in the action fragment;

determining paper product packages and the number of the paper product packages of corresponding specifications based on the current purchased commodities and the purchased commodities;

and performing corresponding production based on the determined specification parameters and the number of the paper product packages.

In one embodiment, the determining the repurchase product of the user according to the purchase information included in the action information and predicting the current repurchase demand degree of the repurchase product includes:

extracting all purchase information about the commodities available for repurchase from the action information of the user on the E-commerce platform;

determining the same commodity which is purchased for multiple times and the corresponding purchasing time thereof from the purchasing information, and estimating the purchasing time interval of the user for the commodity according to the purchasing time;

and judging the re-purchasing demand degree of the user for the commodity according to the latest purchasing time of the user for the commodity and the purchasing time interval.

In one embodiment, the determining, according to the purchase information included in the action information, a re-purchased product of the user and predicting a current re-purchase demand of the re-purchased product further includes:

determining the similar commodities which are purchased for multiple times and the corresponding purchasing time thereof from the purchasing information, and estimating the purchasing time interval of the commodity of the type of the user according to the purchasing time;

and judging the re-purchasing demand degree of the user for the commodities according to the latest time of purchasing the commodities by the user and the purchasing time interval.

In one embodiment, the determining, from the action information, an action fragment of an action of the user to buy the current purchased goods at the current time includes:

respectively acquiring occurrence time intervals of adjacent action occurrence times;

clustering the occurrence time intervals based on density, calculating the mean value of the occurrence time intervals in each class, and taking the occurrence time intervals contained in the class with the maximum mean value as interval nodes;

and dividing the action occurrence time according to the interval nodes to obtain action fragments.

In one embodiment, the predicting the co-purchased product and the demand thereof of the current re-purchased product according to the action information in the action fragment includes:

extracting corresponding commodities in each action fragment to form a plurality of commodity sequences corresponding to the user;

generating a mapping relation between commodities in the commodity sequence in a latest set time period;

and inputting the mapping relation and the corresponding action into a pre-constructed commodity co-purchasing prediction model to predict co-purchasing commodities and co-purchasing demand degrees thereof.

In one embodiment, the determining of the paper product packages and the number thereof with corresponding specifications based on the current purchased products and the purchased products therewith comprises:

the current purchased commodities and the purchased commodities with the current purchased commodities are sequenced together according to respective demand degrees to obtain a demand degree sequence;

extracting a plurality of commodities with highest demand from the demand sequence, so that paper packages required by the commodities are just matched with the production capacity of a paper package production device;

and taking the paper product packages with the corresponding specifications of the plurality of commodities and the quantity thereof as the determined paper product packages and the quantity thereof.

In this document, "first", "second", and the like are used only for distinguishing one from another, and do not indicate their degree of importance, order, and the like.

The division of modules, units or components herein is merely a logical division, and other divisions may be possible in an actual implementation, for example, a plurality of modules and/or units may be combined or integrated in another system. Modules, units, or components described as separate parts may or may not be physically separate. The components displayed as cells may or may not be physical cells, and may be located in a specific place or distributed in grid cells. Therefore, some or all of the units can be selected according to actual needs to implement the scheme of the embodiment.

The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present application should be covered within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims (10)

1. A paper product packaging production management system based on commodity demand prediction is characterized by comprising:

the action information acquisition module is used for acquiring action information of a user on the E-commerce platform;

the repurchase commodity prediction module is used for determining the repurchase commodities of the user according to the purchase information in the action information, predicting the current repurchase demand degree of the repurchase commodities and further predicting the current repurchase commodities of the user;

the co-purchased commodity prediction module is used for determining an action segment of an action of the user for purchasing the current re-purchased commodity in the current date from the action information and predicting the co-purchased commodity and the demand degree of the current re-purchased commodity according to the action information in the action segment;

the paper product package determining module is used for determining paper product packages and the number of the paper product packages in corresponding specifications based on the current purchased commodities and the purchased commodities;

and the paper product package production device is used for carrying out corresponding production based on the determined specification parameters and the number of the paper product packages.

2. The paper product packaging production management system of claim 1, wherein the repurchase goods prediction module comprises:

the purchasing information extraction unit is used for extracting all purchasing information about the commodities available for repeated purchasing from the action information of the user on the E-commerce platform;

the time interval estimation unit is used for determining the same commodity which is purchased for multiple times and the corresponding purchase time thereof from the purchase information and estimating the purchase time interval of the commodity for the user according to the purchase time;

and the repurchase demand judging unit is used for judging the repurchase demand degree of the user for the commodity according to the latest time of purchasing the commodity by the user and the purchasing time interval.

3. The paper product packaging production management system according to claim 2, wherein the time interval estimation unit is further configured to determine similar products purchased multiple times and corresponding purchase times thereof from the purchase information, and estimate a purchase time interval of the user for the similar products based on the purchase times;

the repurchase demand judging unit is also used for judging the repurchase demand degree of the user for the commodities according to the latest time when the user purchases the commodities and the purchase time interval.

4. The paper product packaging production management system of claim 3, wherein the co-purchased goods prediction module comprises:

a time interval acquisition unit for respectively acquiring occurrence time intervals of adjacent action occurrence times;

an interval node acquisition unit, configured to perform density-based clustering on the occurrence time intervals, calculate an average value of the occurrence time intervals in each class, and use the occurrence time intervals included in the class with the largest average value as interval nodes;

and the action fragment dividing unit is used for dividing the action occurrence time according to the interval node to obtain action fragments.

5. The paper product packaging production management system of claim 4, wherein the co-purchased goods prediction module further comprises:

the commodity sequence forming unit is used for extracting corresponding commodities in each action fragment to form a plurality of commodity sequences corresponding to the user;

a mapping relation generation unit for generating a mapping relation between commodities in the commodity sequence in a latest set time period;

and the co-purchase commodity prediction unit is used for inputting the mapping relation and the corresponding action into a pre-constructed commodity co-purchase prediction model and predicting the co-purchase commodities and the co-purchase demand degree.

6. A paper product packaging production management method based on commodity demand prediction is characterized by comprising the following steps:

acquiring action information of a user on an e-commerce platform;

determining the re-purchased commodities of the user according to the purchase information in the action information, predicting the current re-purchased demand degree of the re-purchased commodities, and further predicting the current re-purchased commodities of the user;

determining an action fragment where the action of the user for purchasing the current purchased goods at the current date is located from the action information, and predicting the purchased goods and the demand degree of the current purchased goods according to the action information in the action fragment;

determining paper product packages and the number of the paper product packages of corresponding specifications based on the current purchased commodities and the purchased commodities;

and performing corresponding production based on the determined specification parameters and the number of the paper product packages.

7. The paper product packaging production management method according to claim 6, wherein the determining of the purchased goods of the user and the prediction of the current purchased demand degree of the purchased goods based on the purchase information included in the action information includes:

extracting all purchase information about the commodities available for repurchase from the action information of the user on the E-commerce platform;

determining the same commodity which is purchased for multiple times and the corresponding purchasing time thereof from the purchasing information, and estimating the purchasing time interval of the user for the commodity according to the purchasing time;

and judging the re-purchasing demand degree of the user for the commodity according to the latest purchasing time of the user for the commodity and the purchasing time interval.

8. The paper product packaging production management method according to claim 7, wherein the determining of the purchased goods of the user and the prediction of the current purchased demand degree of the purchased goods based on the purchase information included in the action information further comprises:

determining the similar commodities which are purchased for multiple times and the corresponding purchasing time thereof from the purchasing information, and estimating the purchasing time interval of the commodity of the type of the user according to the purchasing time;

and judging the re-purchasing demand degree of the user for the commodities according to the latest time of purchasing the commodities by the user and the purchasing time interval.

9. The paper product packaging production management method according to claim 8, wherein the determining of the action segment of the action of the user to buy the current repurchase product in the future from the action information includes:

respectively acquiring occurrence time intervals of adjacent action occurrence times;

clustering the occurrence time intervals based on density, calculating the mean value of the occurrence time intervals in each class, and taking the occurrence time intervals contained in the class with the maximum mean value as interval nodes;

and dividing the action occurrence time according to the interval nodes to obtain action fragments.

10. The paper packaging production management method according to claim 9, wherein predicting the co-purchased product and the demand thereof of the currently re-purchased product based on the action information in the action section comprises:

extracting corresponding commodities in each action fragment to form a plurality of commodity sequences corresponding to the user;

generating a mapping relation between commodities in the commodity sequence in a latest set time period;

and inputting the mapping relation and the corresponding action into a pre-constructed commodity co-purchasing prediction model to predict co-purchasing commodities and co-purchasing demand degrees thereof.

Images