CN111814032A

CN111814032A - Cold start recommendation method and device and electronic equipment

Info

Publication number: CN111814032A
Application number: CN201910289260.8A
Authority: CN
Inventors: 苏云琳
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba Group Holding Ltd
Priority date: 2019-04-11
Filing date: 2019-04-11
Publication date: 2020-10-23
Anticipated expiration: 2039-04-11
Also published as: CN111814032B

Abstract

The invention discloses a cold start recommendation method, a cold start recommendation device and electronic equipment, wherein the cold start recommendation method comprises the following steps: according to user data generated by a new user through a third-party application, obtaining a user characteristic value of the new user for setting user characteristics; obtaining a classification characteristic value obtained by classifying each product in a product set according to the user characteristics; according to the user characteristic value and the classification characteristic value of each product, obtaining the matching degree between the new user and each product; and obtaining a product recommendation list of the new user at least according to the matching degree between the new user and each product.

Description

Cold start recommendation method and device and electronic equipment

Technical Field

The present invention relates to the field of information processing technologies, and in particular, to a cold start recommendation method, a cold start recommendation apparatus, an electronic device, and a computer-readable storage medium.

Background

With the development of scientific technology, more and more people tend to read or purchase products, such as reading books, or purchasing commodities, etc., through terminals.

At present, when a user uses a terminal to read or purchase a product for the first time, the user needs to input a large amount of registration information into the terminal, and the terminal recommends a related product to the user according to the large amount of registration information and self equipment information. Taking the example that the user uses the terminal to read the book for the first time, after the user starts the terminal for the first time, the corresponding table is displayed on the display interface of the terminal to prompt the user to input a large amount of registration information such as the sex of the user and the type of the book of interest. Meanwhile, the terminal actively acquires the ip address of the device per se and inquires the region information of the user according to the ip address. And finally, the terminal displays the books recommended to the user on a display interface according to the registration information and the region information of the user.

However, on the one hand, the user is required to input a large amount of registration information into the terminal, which results in cumbersome user operation. On the other hand, the terminal cannot obtain the characteristics that accurately reflect the relevant information of the user according to the device information of the terminal, for example, the information of the region where the user is located cannot be accurately inquired according to the ip address of the device of the terminal, and meanwhile, due to the reason of complicated operation, the user does not necessarily input all the registration information, so that the product recommended by the terminal to the user cannot meet the actual requirements of the user.

Disclosure of Invention

The embodiment of the invention aims to provide a new cold start recommendation method for recommending products for a new registered user.

According to a first aspect of the present invention, there is provided a cold start recommendation method, comprising:

according to user data generated by a new user through a third-party application, obtaining a user characteristic value of the new user for setting user characteristics;

obtaining a classification characteristic value obtained by classifying each product in a product set according to the user characteristics;

according to the user characteristic value and the classification characteristic value of each product, obtaining the matching degree between the new user and each product;

and obtaining a product recommendation list of the new user at least according to the matching degree between the new user and each product.

Optionally, the user characteristics include at least one of a gender characteristic, an age characteristic, a geographic characteristic, and a scholarly calendar characteristic.

Optionally, the obtaining, according to the user feature value and the classification feature value of each product, a matching degree between the new user and each product includes:

according to the classification characteristic value of each product, acquiring the preference degree of the corresponding product for each user characteristic value;

and obtaining the matching degree between the new user and each product according to the preference degree corresponding to each product and the influence weight of each user characteristic.

Optionally, wherein the method further comprises:

and acquiring the search heat of each product in a search engine so as to obtain a product recommendation list of the new user at least according to the search heat.

Optionally, wherein the method further comprises:

and acquiring the transaction heat formed by each product through transaction payment so as to obtain a product recommendation list of the new user at least according to the transaction heat.

Optionally, wherein the method further comprises:

comparing the user characteristic value of the new user with the user characteristic value of the historical new user to obtain a comparison result;

and obtaining the degree of cooperation of each product according to the comparison result and the products converted by the historical new user within the set time after registration, so as to obtain a product recommendation list of the new user at least according to the degree of cooperation, wherein the converted products belong to the product set.

Optionally, the obtaining the synergy of each product according to the comparison result and the product converted by the historical new user within the set time after registration includes:

acquiring a product converted by the historical new user within a set time after registration;

obtaining a classification characteristic value obtained by classifying each converted product in the historical new user according to the user characteristics;

and obtaining the degree of cooperation of each product according to the comparison result and the classification characteristic value of each converted product.

Optionally, after obtaining the product recommendation list of the new user, the method further includes:

obtaining a product list actually converted by the new user;

and comparing the actually converted product list with the product recommendation list, and updating the book recommendation list according to the comparison result.

Optionally, wherein the method further comprises:

screening products to be recommended from all the products according to set conditions to form a product set;

wherein the set condition comprises at least one of the search heat in the search engine exceeding the set search heat, the transaction heat formed by the transaction payment exceeding the set transaction heat, and the conversion rate in the historical new user exceeding the set conversion rate.

According to a second aspect of the present invention, there is provided a cold start recommendation method, comprising:

obtaining a classification characteristic value obtained by classifying each book in the book set according to the user characteristics;

obtaining the matching degree between the new user and each book according to the user characteristic value and the classification characteristic value of each book;

and obtaining a book recommendation list of the new user at least according to the matching degree between the new user and each book.

Optionally, wherein the method further comprises:

acquiring the search heat of each product in a search engine;

acquiring transaction heat generated by each product through transaction payment;

obtaining the degree of cooperation of each product according to the comparison result and the products converted by the historical new user within the set time after registration, wherein the converted products belong to the product set;

the method also obtains a product recommendation list of the new user according to the search heat, the transaction heat and the cooperation of each product.

According to a third aspect of the present invention, there is provided a cold start recommendation device comprising:

the user characteristic acquisition module is used for acquiring a user characteristic value of a new user for a set user characteristic according to user data generated by the new user through the third-party application;

the product characteristic acquisition module is used for acquiring a classification characteristic value obtained by classifying each product in the product set according to the user characteristic;

the matching module is used for obtaining the matching degree between the new user and each product according to the user characteristic value and the classification characteristic value of each product; and the number of the first and second groups,

and the recommending module is used for obtaining a product recommending list of the new user at least according to the matching degree between the new user and each product.

According to a fourth aspect of the present invention, there is provided an electronic device comprising the cold start recommendation apparatus according to the third aspect of the present invention; or the electronic device comprises a memory for storing computer instructions and a processor for calling the computer instructions from the memory and executing the cold start recommendation method according to the first aspect or the second aspect under the control of the computer instructions.

According to a fifth aspect of the present invention, there is provided a computer storage medium storing computer instructions which, when executed by a processor, implement the cold start recommendation method according to the first or second aspect.

The embodiment of the invention provides a cold start recommendation method, and on one hand, a product recommendation list can be obtained without inputting a large amount of registration information by a new user, so that the relevant operation of inputting the registration information by the user can be avoided. On the other hand, the matching degree between the new user and each product can be obtained according to the accurate and rich characteristic values representing the user characteristics and the classification characteristic values of the products classified according to the user characteristics, which are obtained by the new user through the user data generated by using the third-party application, and the product recommendation list close to the actual requirements of the new user is obtained at least according to the matching degree between the new user and each product.

Other features of the present invention and advantages thereof will become apparent from the following detailed description of exemplary embodiments thereof, which proceeds with reference to the accompanying drawings.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required to be used in the embodiments will be briefly described below. It is appreciated that the following drawings depict only certain embodiments of the invention and are therefore not to be considered limiting of its scope. For a person skilled in the art, it is possible to derive other relevant figures from these figures without inventive effort.

FIG. 1 is a schematic diagram of a product providing system to which the cold start recommendation method of an embodiment of the present invention can be applied;

FIG. 2 is a flowchart illustrating a cold start recommendation method according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of a product recommendation list according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of another product recommendation list provided by the embodiment of the present invention in a display form;

FIG. 5 is a flowchart illustrating another method for cold start recommendation according to an embodiment of the present invention;

FIG. 6 is a flowchart illustrating a method for cold start recommendation according to another embodiment of the present invention;

FIG. 7 is a flowchart illustrating a method for cold start recommendation according to an embodiment of the present invention;

fig. 8 is a schematic structural diagram of a cold start recommendation device according to an embodiment of the present invention.

Detailed Description

Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that: the relative arrangement of the components and steps, the numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present invention unless specifically stated otherwise.

The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail but are intended to be part of the specification where appropriate.

In all examples shown and discussed herein, any particular value should be construed as merely illustrative, and not limiting. Thus, other examples of the exemplary embodiments may have different values.

It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, further discussion thereof is not required in subsequent figures.

< hardware configuration >

The cold start recommendation method provided by the embodiment of the invention can be applied to a product providing system, and the product providing system can recommend products for new users according to the cold start recommendation method provided by the embodiment of the invention, so that the conversion rate of the users to the recommended products is improved, and further the user experience is improved. Fig. 1 is a schematic structural diagram of a product providing system to which the cold start recommendation method according to the embodiment of the present invention can be applied.

As shown in fig. 1, such a product providing system 1000 may include a server 1100, a terminal apparatus 1200, and a network 1300.

The server 1100 may be, for example, a blade server, a rack server, etc., and is not limited herein.

As shown in FIG. 1, server 1100 may include a processor 1110, a memory 1120, an interface device 1130, a communication device 1140, a display device 1150, and an input device 1160.

The processor 1110 may be, for example, a central processing unit CPU or the like. The memory 1120 includes, for example, a ROM (read only memory), a RAM (random access memory), a nonvolatile memory such as a hard disk, and the like. The interface device 1130 includes, for example, a USB interface, a serial interface, and the like. The communication device 1140 is capable of wired or wireless communication, for example. The display device 1150 is, for example, a liquid crystal display panel. Input devices 1160 may include, for example, a touch screen, a keyboard, and the like.

Although shown as multiple devices in fig. 1, server 1100 may also include only processor 1110, memory 1120, and communications device 1140.

In one embodiment, the servers may be implemented as a cloud architecture, e.g., by a cluster of servers deployed in the cloud.

The terminal device 1200 may be a smartphone, laptop, desktop, tablet, wearable device, PDA, e-reader, etc. The terminal device 1200 may be any device capable of loading a product providing application, for example, a device capable of providing a service such as online or offline reading of literary works, and the like, which is not limited herein.

As shown in fig. 1, the terminal apparatus 1200 may include a processor 1210, a memory 1220, an interface device 1230, a communication device 1240, a display device 1250, an input device 1260, a speaker 1270, a microphone 1280, and the like. The processor 1210 may be a central processing unit CPU, a microprocessor MCU, or the like. The memory 1220 includes, for example, a ROM (read only memory), a RAM (random access memory), a nonvolatile memory such as a hard disk, and the like. The interface device 1230 includes, for example, a USB interface, a headphone interface, and the like. The communication device 1240 can perform wired or wireless communication, for example. The display device 1250 is, for example, a liquid crystal display, a touch display, or the like. The input device 1260 may include, for example, a touch screen, a keyboard, and the like. A user can output/input voice information through the speaker 1270 and the microphone 1280.

The communication network 1300 may be a wireless network or a wired network, and may be a local area network or a wide area network. The terminal apparatus 1200 can communicate with the server 1100 through the communication network 1300.

The product delivery system 1000 shown in FIG. 1 is illustrative only and is not intended to limit the invention, its application, or uses in any way. For example, although fig. 1 shows only one server 1100 and one terminal device 1200, it is not meant to limit the respective numbers, and the product providing system 1000 to which the cold start recommendation method of the embodiment of the present invention is applied may include a plurality of servers 1100 and/or a plurality of terminal devices 1200.

The cold start recommendation method according to any embodiment of the present invention may be implemented by the server 1100 of the product providing system 1000, may be implemented by the terminal device 1200 of the product providing system 1000, and may also be implemented by both the server 1100 and the terminal device 1200, which is not limited herein.

< example one >

The cold start recommendation method provided in this embodiment, as shown in fig. 2, includes the following steps S101 to S104:

s101, obtaining a user characteristic value of the new user for setting user characteristics according to user data generated by the new user through the third-party application.

In this embodiment, the new user may be a user whose current date is within a set time period from the registration date, for example, if the set time period is one month, all users whose current date is within one month from the registration date are called new users.

In this embodiment, the third-party application may include: at least one of a video playing platform, a search engine, a payment platform, a navigation positioning platform and an online shopping platform.

The user data generated by the new user through the third-party application may be, for example:

when the third-party application is a video playing platform, the corresponding data may include: the registration information of the user, the type of videos the user has historically watched (e.g., comedy, love, home, city, idol, news, documentary, etc.), the duration of the videos the user has historically watched, the corresponding region of the videos the user has historically watched (e.g., inland, hong kong, europe, america, japan, korea, etc.), etc.

When the third party application is a search engine, the corresponding data may include: registration information of the user, user history search records, and the like.

When the third party application is for a payment platform, the corresponding data may include: registration information for the user, the user's historical payment bill, the type of merchandise paid (e.g., dining, clothing, daily chemicals, etc.), and the like.

When the third party application is for navigating a positioning platform, the corresponding data may include: the registered information of the user, the origin of the user's historical trip, the destination of the user's historical trip, the user's historical trip pattern (e.g., walking, taxi, riding, driving, etc.), the type of the surrounding building (office building, mall, vegetable market, etc.) of the origin and destination of the user's historical trip, etc.

When the third-party application is an online shopping platform, the corresponding data may include: the system comprises registration information of a user, historical order information of the user (the order information comprises a receiving address, a commodity type, a commodity price and the like), commodity information browsed by the user in history and the like.

Illustratively, the user characteristics may include: at least one of a gender feature, an age feature, a geographic feature, and a scholarly calendar feature.

The characteristic values of the gender characteristics are: male or female. The characteristic value of the age characteristic may be divided according to age groups, for example, into: under 20 years old, 20-35 years old, 35-50 years old, over 50 years old, and the like. The characteristic value of the region characteristic can be divided according to provinces, can also be divided according to cities, and the like, for example, the characteristic value is divided into: beijing, Shanghai, Guangzhou, Shaanxi province, Jiangsu province, Ningxia, Taiwan province, etc. The feature values of the academic characters can be divided into: junior middle school and the following school calendars, high school calendars, university calendars, student calendars and the like.

It should be noted that the feature value of the user feature may be further quantized. For example, the characteristic values of the sex characteristics are quantified as 0 and 1, where 0 represents male and 1 represents female.

In addition, the user characteristics may further include type characteristics of the type of interest, and the characteristic value of the type characteristics of the type of interest may be: comedy, love, family, city, youth idol, news, documentary, etc., without limitation.

In this embodiment, when the step S101 is implemented, the product providing system 1000 applying the cold start recommendation method according to the embodiment of the present invention may first obtain device related information of the new user, for example, obtain an identification number of the device of the new user, and registration information provided by the new user when the product providing system 1000 registers, so as to search the new user in the third-party application according to the information, and further obtain user data generated by the new user through using the third-party application. For example, the product providing system 1000 may communicate data with a third party platform through International Mobile Equipment Identity (IMEI) and UTD-ID of a handset used by a new user.

In an example, the implementation manner of the step S101 may include: and determining the characteristic value of the regional characteristic of the new user according to the origin of the historical trip, the destination of the historical trip of the user, the types of buildings around the origin and the destination of the historical trip of the user and/or the receiving address of the order in the network management platform, wherein the origins and the destinations are provided by the navigation positioning platform. For example, the province where the most frequent buildings are the origins of the stores and/or the most frequent receiving addresses are used in the historical trips is determined as the characteristic value of the regional characteristic of the new user.

In an example, the implementation manner of the step S101 may include: and determining the characteristic value of the academic character of the new user according to the historical search records of the user provided by the search engine. For example, the search content of the historical search record of the user is determined, the proportion of related teaching materials and/or related knowledge in different academic stages included in the search content is determined, the academic records corresponding to the teaching materials and/or knowledge with the highest proportion are determined, and the characteristic value of the academic character of the new user is determined.

In an example, the implementation manner of the step S101 may include: and determining the age characteristics of the new user according to the video types watched by the user history provided by the video playing platform. For example, an optimum viewing age range corresponding to the type of video viewed in the history is determined, and the optimum viewing age range is used as a feature value of the age feature of the new user.

In an example, the implementation manner of the step S101 may include: and determining the characteristic value of the gender characteristic of the new user according to the paid commodity type provided by the payment platform. For example, when the type of the paid commodity is determined as clothing, the ratio of male clothing to female clothing in the clothing is determined, and the corresponding gender of the clothing with the higher ratio is used as the characteristic value of the gender characteristic.

S102, obtaining a classification characteristic value obtained by classifying each product in the product set according to the user characteristics.

The above products may be, for example, books, electronic products, cosmetics, clothes, and the like, and are not limited herein.

In one example, the product set described above may include a set of all products that the product providing system 1000 is capable of providing.

In another example, the product set may be a set of products selected from all products according to a set condition, so that the information processing amount is reduced and the recommendation efficiency is improved. Based on this, the cold start recommendation method provided in this embodiment may further include: and screening products to be recommended from all the products according to set conditions to form the product set.

In this example, the above-mentioned setting condition may include at least one of a search heat exceeding a set search heat in a search engine, a transaction heat exceeding a set transaction heat formed by a transaction payment, and a conversion rate exceeding a set conversion rate in a history of new users.

The search heat may be quantified, for example, according to the number of times the product is searched within a preset time period.

The transaction popularity may be quantified, for example, based on the number of times the product was successfully traded within a predetermined time period. In the case of books, the transaction payment may include at least one of a reading payment and a purchase payment.

Taking a product as a product a as an example, for the product a, the conversion rate in the historical new user may be: and aiming at all the historical new users containing the product A in the recommended product, converting the ratio of the number of the people of the product A to the total number of the historical new users containing the product A in the recommended product. The meaning of the above conversion can be determined according to the need and the type of the provided product, for example, the product providing system 1000 provides an electronic book reading service, and the conversion can refer to a book whose reading progress reaches the set condition, and the reading progress can include at least one of reading amount and reading duration; as another example, if product-providing system 1000 provides a service that sells a certain type of product, then a conversion may refer to a successful sale, and so forth.

Taking the user characteristics including age characteristics, gender characteristics, academic history characteristics, and regional characteristics as an example, the classification characteristic value obtained by classifying each product in the product set according to the user characteristics in step S102 may be represented as a1-a4 as follows:

and a1, obtaining the classification characteristic value of each product in the product set according to the gender characteristic classification.

For example, for each product in the product set, the number of female users who converted the product X1, the number of male users who converted the product X2, the number of female users among all users Y1, and the number of male users among all users Y2 are obtained.

This may be directly using X1 and Y1, and X2 and Y2 as classification feature values obtained by gender classification.

This may be achieved by directly using X1 and X2 as classification feature values classified by gender.

This may be achieved by using the preference of the product for women and the preference for men as classification feature values classified by gender.

Preference profile pre of each product for any classification feature j _ i of any user feature j_{j_i}It can be calculated according to the following formula (1):

in formula (1), j _ i represents the serial number of the classification features obtained by classifying the provided product according to the user feature j, where j is the serial number of the user feature, for example, j ═ 1 represents a gender feature, j ═ 2 represents an age feature, j ═ 3 represents a academic calendar feature, j ═ 4 represents a regional feature, and the user feature is taken as a gender feature as an example, the classification features include two classification features, for example, i ═ 1 represents a woman, i ═ 2 represents a man, and then profile _ pre_{1_1}Prifile _ pre, representing the preference of a product for women_{1_2}Representing the preference of the product for men. n represents the number of classification features obtained by classifying the provided product according to a user feature j, for example, for classification according to gender, the obtained classification features include female and male, so the number of classification features obtained by classification according to gender is 2, n is 2, for classification according to age, the obtained classification features are four, so the number of classification features obtained by classification according to age is 4, n is 4. Read _ uv_{j_i}Representing the number of people who converted a product that meets the classification characteristic of serial number j _ i. Read _ uv _ all_{j_i}Represents the number of persons who meet the classification characteristic of the serial number j _ i among all the users of the product providing system 1000.

In this example, for the user characteristic being a gender characteristic, the classification characteristic includes a female (denoted by the number 1) and a male (denoted by the number 2), and the preference of a product for the female and the preference of the product for the male can be calculated according to the formula (1), that is:

male preference (X2/X1)/(Y2/Y1);

women preference (X1/X2)/(Y1/Y2).

a2, obtaining the characteristic value of the age characteristic of each product in the product set according to the age characteristic classification.

For example, for each product in the product set, the number Z1 of users who transform the product and are under the age of 20, the number Z2 of users who transform the product and are under the age of 20 to 35, the number Z3 of users who transform the product and are under the age of 35 to 50, the number Z4 of users who transform the product and are over the age of 50, the number W1 of users under the age of 20 among all users, the number W2 of users under the age of 20 to 35 among all users, the number W3 of users under the age of 35 to 50 among all users, and the number W4 of users over the age of 50 among all users are obtained.

This may be achieved by directly using Z1 and W1, Z2 and W2, Z3 and W3, and Z4 and W4 as classification feature values obtained by classifying according to age.

This may be achieved by directly using Z1, Z2, Z3, and Z4 as classification feature values classified by age.

Similarly, this may be to use the preference degrees of the product for different age categories as the classification feature values obtained by gender classification.

Still referring to the above formula (1), the user features are age features, and the classification features thereof may include, in terms of age groups, 20 years or less (denoted by reference numeral 1), 20 years to 35 years (denoted by reference numeral 2), 35 to 50 years (denoted by reference numeral 3), and 50 years or more (denoted by reference numeral 4), that is, the age features have 4 classification features, and n is 4 in the corresponding formula (1), and the preference degrees of a product with respect to the 4 classification features included in the age features may be calculated according to the formula (1) as:

a3, obtaining the classification characteristic value of the academic record characteristic, wherein the academic record characteristic is obtained by classifying each product in the product set according to the academic record characteristic.

For example, for each product in the product set, the number P1 of users who convert the product into the academic calendar of junior middle school and below, the number P2 of users who convert the product into the academic calendar of high school, the number P3 of users who convert the product into the academic calendar of university, the number P4 of users who convert the academic calendar of the product into the academic calendar of students, the number Q1 of users who have the academic calendar of junior middle school and below in all users, the number Q2 of users who have the academic calendar of high school in all users, the number Q3 of users who have the academic calendar of university in all users, and the number Q4 of users who have the academic calendar of research students in all users are obtained.

This may be done by directly using P1 and Q1, P2 and Q2, P3 and Q3, and P4 and Q4 as classification feature values obtained by classification according to academic calendars.

This may be achieved by directly using P1, P2, P3, and P4 as classification feature values classified according to the academic calendar.

Similarly, this may be to use the preference degrees of the product for different academic categories as the classification feature values obtained by the academic categories.

Still referring to the above formula (1), the user features are academic records, and the classification features thereof may include a junior middle school calendar and the following academic records (denoted by the reference numeral 1), a high school calendar (denoted by the reference numeral 2), a university academic record (denoted by the reference numeral 3), a student academic record (denoted by the reference numeral 4), and the like, that is, the academic records features have 4 classification features, and the preference of a product with respect to the above 4 classification features included in the academic records features may be calculated according to the formula (1).

a4, obtaining the classification characteristic value of the region characteristic of each product in the product set according to the region characteristic classification.

Similarly, for the regional characteristics, the classification characteristic values obtained by classifying a product according to the academic history can be obtained by referring to the above examples a1 to a3, and the description thereof is omitted.

S103, obtaining the matching degree between the new user and each product according to the user characteristic value and the classification characteristic value of each product.

Continuing to take the user features as gender features, age features, geographic features and academic features as examples, in one example, in S103, the matching degree between the new user and each product is obtained according to the user feature value and the classification feature value of each product, which can be implemented by the following steps S1031 and S1032:

and S1031, acquiring the preference degree of the corresponding product for each user characteristic value according to the classification characteristic value of each product.

Taking the user characteristic as the gender characteristic as an example, if the characteristic value of the gender characteristic of the new user is female, the preference degree of each product for female can be directly obtained from the classification characteristic value of the product according to different forms of the classification characteristic value, or the preference degree of each product for female can be obtained according to the classification characteristic value of the product and the formula (1).

Taking the user characteristic as the age characteristic as an example, if the characteristic value of the age characteristic of the new user is below 20 years old, the preference of each product for the age group below 20 years old can be directly obtained from the classification characteristic value of the product according to different forms of the classification characteristic value, or the preference of each product for the age group below 20 years old can be obtained according to the classification characteristic value of the product and the above formula (1).

S1032, obtaining the matching degree between the new user and each product according to the preference degree corresponding to each product and the influence weight of each user characteristic.

For example, the above expression for calculating the matching degree M may be as the following formula (2):

in the formula (2), M indicates a degree of matching. j represents the serial number of the user feature, for example, in the embodiment of the present invention, four user features are selected and respectively include a gender feature, an age feature, a study history feature, and a region feature, and then the serial number of the gender feature is 1, the serial number of the age feature is 2, the serial number of the study history feature is 3, and the serial number of the region feature is 4. w is a_jRefers to the weight of influence of the user's feature with sequence number j, where w_jThe influence weights of different user characteristics may be the same or different, and are not limited hereinE.g. w₁Representing the influence weight of the gender characteristics. m is the number of user features, for example, if the above four user features are selected, m is 4. In formula (2), j _ i is determined according to the user feature value of the new user for the user feature j, for example, if the feature value of the new user for the gender feature is female, j _ i is 1_1, for example, if the feature value of the new user for the age feature is 20 years old or less, j _ i is 2_1, for example, if the feature value of the new user for the academic calendar feature is the research student calendar, j _ i is 3_4, and so on, and details are not repeated.

The user characteristic value of each user characteristic of the new user comprises: for example, women, 20 years old or younger, the academic calendar of the research student, and the northeast region, according to the above formula (2), the matching degree between the new user and each product can be obtained according to the preference degree of each product for each user feature value and the influence weight of the corresponding user feature.

And S104, obtaining a product recommendation list of the new user at least according to the matching degree between the new user and each product.

The implementation manner of the step S104 may be: based on the matching degree between the new user and each product obtained in the step S103, N products with the highest matching degree are extracted as recommended products, and a product recommendation list is generated based on the related information of the recommended products. Further, the product recommendation list may be presented to the new user.

For example, taking a product as a book, the related information of the recommended product may be: book name, author, publisher, book cover image, etc.

For example, the above-mentioned product recommendation list may be displayed in the form shown in fig. 3 and fig. 4.

< example two >

As shown in fig. 5, the cold start recommendation method provided in this embodiment may include the following steps S501 to S505:

s501, obtaining a user characteristic value of the new user for the set user characteristic according to user data generated by the new user through the third-party application.

S502, obtaining a classification characteristic value obtained by classifying each product in the product set according to the user characteristics.

S503, obtaining the matching degree between the new user and each product according to the user characteristic value and the classification characteristic value of each product.

It should be noted that the specific implementations of the above-mentioned S501-S503 are the same as the specific implementations of the above-mentioned S101-S103, and are not described herein again.

S504, obtaining the searching heat of each product in the search engine.

In this embodiment, the search heat may be quantified according to the number of times the product is searched within a preset time period, for example.

And S505, obtaining a product recommendation list of the new user at least according to the matching degree and the searching popularity between the new user and each product.

In this embodiment, since the search popularity of the product in the search engine is also a factor that affects the conversion of the product by the new user, the recommendation list of the product may be further determined according to the search popularity of the product in the search engine.

Based on the above, when the search heat is quantitatively expressed by H, the product recommendation list of the new user can also be obtained based on the output value f1 of the following formula (3):

wherein, K1, K2 represent the weight of the matching degree and the searching heat degree respectively, K1, K2 are both larger than 0, and the sum is 1. The description of the other parameters in the above formula (3) is the same as the description of the parameters in the above formula (2), and is not repeated here.

In this embodiment, based on f1 obtained by the above formula (3), the highest N products corresponding to f1 are extracted as recommended products, and a product recommendation list is generated based on the related information of the recommended products. Further, the product recommendation list may be presented to the new user.

For example, the above-mentioned product recommendation list may also be displayed in the form shown in fig. 3 and fig. 4.

Based on the embodiment, when the product recommendation list of the new user is obtained, the factor of the search heat which can influence the new user to convert the product is considered, so that the finally obtained recommendation list is closer to the actual demand of the user.

< example three >

As shown in fig. 6, the cold start recommendation method provided in this embodiment may include the following steps S601-S605:

s601, according to user data generated by the new user through the third-party application, obtaining a user characteristic value of the new user for the set user characteristic.

S602, obtaining a classification characteristic value obtained by classifying each product in the product set according to the user characteristics.

And S603, obtaining the matching degree between the new user and each product according to the user characteristic value and the classification characteristic value of each product.

It should be noted that the specific implementations of the above S601 to S603 are the same as the specific implementations of the above S101 to S103, and are not described again here.

And S604, acquiring the transaction heat formed by each product through transaction payment.

In this embodiment, the transaction popularity may be quantified according to the number of times that the product is successfully traded within a predetermined time period.

S605, obtaining a product recommendation list of the new user at least according to the matching degree and the transaction popularity between the new user and each product.

In this embodiment, since the transaction popularity of the product in the transaction is also a factor that affects the conversion of the product by the new user, the recommendation list of the product may be further determined according to the transaction popularity of the product in the transaction payment.

Based on the above, when the transaction popularity is represented by P, the product recommendation list of the new user can also be obtained based on the output value f2 of the following formula (4):

wherein, K3, K4 represent the weight of the matching degree and the transaction popularity respectively, K3, K4 are both greater than 0, and the sum is 1. The description of the other parameters in the above formula (4) is the same as the description of the parameters in the above formula (2), and is not repeated here.

In one embodiment, when the transaction popularity is represented by P, the product recommendation list of the new user can also be obtained based on the output value f2 as shown in the following formula (5):

wherein, K5, K6, K7 respectively represent the weight of the matching degree, the search degree and the transaction degree, K5, K6 and K7 are all larger than 0, and the sum is 1. The description of the other parameters in the above formula (5) is the same as the description of the parameters in the above formula (2) and the above formula (3), and will not be described again here.

Based on the above, in the cold start recommendation method provided in this embodiment, the step S605 may be replaced by: and obtaining a product recommendation list of the new user at least according to the matching degree, the transaction popularity and the search popularity between the new user and each product.

In this embodiment, based on f2 obtained by the above formula (4) or formula (5), the highest N products corresponding to f2 are extracted as recommended products, and a product recommendation list is generated based on the related information of the recommended products. Further, the product recommendation list may be presented to the new user.

Based on the embodiment, when the product recommendation list of the new user is obtained, the factor of transaction heat which can influence the conversion of the product by the new user is considered, so that the finally obtained recommendation list is closer to the actual demand of the user.

< example four >

As shown in fig. 7, the cold start recommendation method provided in this embodiment may include the following steps S701 to S706:

s701, according to user data generated by the new user through the third-party application, obtaining a user characteristic value of the new user for the set user characteristic.

S702, obtaining a classification characteristic value obtained by classifying each product in the product set according to the user characteristics.

And S703, obtaining the matching degree between the new user and each product according to the user characteristic value and the classification characteristic value of each product.

It should be noted that the specific implementation of S701 to S703 is the same as the specific implementation of S101 to S103, and is not described herein again.

S704, comparing the user characteristic value of the new user with the user characteristic value of the historical new user to obtain a comparison result.

In this embodiment, the specific implementation manner of S704 is: and according to the user characteristic value of the new user, based on a collaborative filtering algorithm, finding the user characteristic value of the historical new user which is most similar to the user characteristic value of the new user from the user characteristic values of the historical new users. And using the user characteristic value of the most similar historical new user as a comparison result.

S705, obtaining the degree of cooperation of each product according to the comparison result and the products converted by the new historical user within the set time after registration.

In one embodiment, the above S705 can be implemented by the following steps S7051 to S7053:

s7051, obtaining products converted by the new historical user within the set time after registration.

In the present embodiment, the history new user refers to a user whose current date is more than a preset time length from the registration date.

S7052, obtaining a classification characteristic value of each conversion product in the historical new users according to the product characteristic classification.

In the present embodiment, the specific implementation of S7052 is similar to that of S102, except that when S7052 is executed, it is determined based on the historical new user.

For example, when the product features are used as the gender features, the classification feature values obtained by classifying the conversion products in the historical new users according to the gender features may be: acquiring the number of female users in the historical new users for converting the product, the number of male users in the historical new users for converting the product, the number of female users in all the historical new users and the number of male users in all the historical new users.

S7053, obtaining the degree of coordination of each product according to the comparison result and the classification characteristic value of each converted product.

In this embodiment, for one converted product, the classification feature value of the one converted product, and the comparison result obtained based on the above S704 are: and (3) substituting the user characteristic value of the most similar historical new user into the calculation formula of the matching degree in the formula (2), and taking the finally obtained matching degree as the cooperation degree of the converted product. By repeating this step, the degree of synergy of each converted product can be obtained.

In another embodiment, the step of S705 above may be implemented as follows: based on the comparison result obtained in S704 above: determining the user characteristic value of the most similar historical new user and the corresponding product converted by the historical new user;

for each of the products converted by the historical new user, the classification characteristic value of the one converted product and the comparison result obtained based on the above step S704 are: and (3) substituting the user characteristic value of the most similar historical new user into the calculation formula of the matching degree in the formula (2), and taking the finally obtained matching degree as the cooperation degree of the converted product. By repeating this step, the degree of synergy of each converted product can be obtained.

S706, obtaining a product recommendation list of the new user at least according to the matching degree and the cooperation degree between the new user and each product, wherein the converted products belong to a product set.

In this embodiment, since the degree of synergy of the user feature values of the historical new users, which are most similar to the user feature value of the new user, of the product pair converted by the historical new user is also a factor that affects the conversion of the product by the new user, the recommendation list of the product may be further determined according to the degree of synergy.

Based on the above, when the degree of cooperation is represented by C, the product recommendation list of the new user can also be obtained based on the output value f3 of the following formula (6):

wherein, K8 and K9 respectively represent the weight of the matching degree and the cooperation degree, K8 and K9 are both larger than 0, and the sum is 1. The description of the other parameters in the above formula (6) is the same as the description of the parameters in the above formula (2), and is not repeated here.

In one embodiment, when the degree of synergy is represented by C, the product recommendation list of the new user may also be obtained based on the output value f3 as the following formula (7):

wherein, K10, K11 and K12 respectively represent the weight of the matching degree, the transaction heat degree and the cooperation degree, K10, K11 and K12 are all larger than 0, and the sum is 1. The description of the other parameters in the above formula (7) is the same as the description of the parameters in the above formula (2) and formula (4), and is not repeated here.

Based on the above, in the cold start recommendation method provided in this embodiment, the step S706 may be replaced by: and obtaining a product recommendation list of the new user at least according to the matching degree, the cooperation degree and the transaction heat degree between the new user and each product, wherein the converted products belong to a product set.

In one embodiment, when the degree of synergy is represented by C, the product recommendation list of the new user may also be obtained based on the output value f3 as the following formula (8):

wherein, K13, K14, K15 respectively represent the weight of the matching degree, the search heat degree and the cooperation degree, K13, K14, K15 are all larger than 0, and the sum is 1. The description of the other parameters in the above formula (8) is the same as the description of the parameters in the above formula (2) and formula (3), and is not repeated here.

Based on the above, in the cold start recommendation method provided in this embodiment, the step S706 may be replaced by: and obtaining a product recommendation list of the new user at least according to the matching degree, the cooperation degree and the search heat degree between the new user and each product, wherein the converted products belong to a product set.

In one embodiment, when the degree of synergy is represented by C, the product recommendation list of the new user may also be obtained based on the output value f3 as the following formula (9):

wherein K16, K17, K18 and K19 respectively represent the weight of the matching degree, the search heat degree, the cooperation degree and the transaction heat degree, and K16, K17, K18 and K19 are all greater than 0 and the sum is 1. The description of the other parameters in the above formula (9) is the same as the description of the parameters in the above formula (2), formula (3), and formula (4), and will not be repeated here.

Based on the above, in the cold start recommendation method provided in this embodiment, the step S706 may be replaced by: and obtaining a product recommendation list of the new user at least according to the matching degree, the cooperation degree, the search heat degree and the transaction heat degree between the new user and each product, wherein the converted product belongs to a product set.

It should be noted that, when the converted product corresponding to the cooperation degree C is different from the product corresponding to the matching degree M, the cooperation degree C may be exemplarily set to 0.

In this embodiment, based on f3 obtained from the above formulas (6) to (8), the highest N products corresponding to f3 are extracted as recommended products, and a product recommendation list is generated based on the related information of the recommended products. Further, the product recommendation list may be presented to the new user.

Based on the embodiment, when the product recommendation list of the new user is obtained, the factor of the synergy which can influence the conversion of the products of the historical new user is considered, so that the finally obtained recommendation list is closer to the actual demand of the user.

< example five >

On the basis of any one of the above embodiments, the cold start recommendation method provided by the present invention further includes the following steps S105 and S106:

and S105, acquiring a product list actually converted by the new user.

Specifically, the product list actually converted by the new user is a list formed by products actually converted by the new user within a preset time period. Here, the preset time period may be 10 days or the like.

And S106, comparing the actually converted product list with the product recommendation list, and updating the book recommendation list according to the comparison result.

In this embodiment, when the above S106 is executed, the number of products included in both the product list and the product recommendation list actually converted by the new user may be determined first. Then, according to the ratio of the number to the total number of the products in the product recommendation list, determining whether to adjust and calculate w in the matching degree formula_j。

If the ratio is larger than the preset ratio, not comparing w in the matching degree formula_iAnd (6) adjusting. On the contrary, for w in the matching degree formula in the above S1032_jAnd (6) adjusting.

Illustratively, for w in the above matching degree formula_jThe adjustment method may be: determining the preference degree with large integral value in the preference degrees corresponding to the products contained in the product list and the product recommendation list of the actual conversion of the new user, and corresponding the preference degree to w_jThe value of (c) is increased. Then based on the adjusted w_jAnd obtaining a subsequent product recommendation list.

Based on the above embodiments, in the cold start recommendation method provided by the embodiments of the present invention, the obtained book recommendation list may be updated, so that the book recommendation list obtained by the subsequent new user is closer to the actual requirement of the subsequent new user.

< example six >

The embodiment of the invention provides a cold start recommendation method for books, which comprises the following steps of S201-S204:

s201, according to user data generated by the new user through the third-party application, obtaining a user characteristic value of the new user for the set user characteristic.

S202, obtaining a classification characteristic value obtained by classifying each book in the book set according to the user characteristics.

S203, obtaining the matching degree between the new user and each book according to the user characteristic value and the classification characteristic value of each book.

And S204, obtaining a book recommendation list of the new user at least according to the matching degree between the new user and each book.

In an embodiment, an embodiment of the present invention provides a cold start recommendation method, further including the following S205-S204:

s205, obtaining the search heat of each product in the search engine.

And S206, acquiring the transaction heat generated by each product through transaction payment.

And S207, comparing the user characteristic value of the new user with the user characteristic value of the historical new user to obtain a comparison result.

And S208, obtaining the degree of coordination of each product according to the comparison result and the converted products of the new historical users within the set time after registration, wherein the converted products belong to a product set.

S209, the method also obtains a product recommendation list of the new user according to the search heat, the transaction heat and the cooperation of each product.

It should be noted that the book provided in the embodiment of the present invention may be used as a product related to any one of the cold start recommendation methods in the first to fifth embodiments.

The above method embodiments are described in a progressive manner, and the same and similar parts among the various embodiments may be referred to each other, each embodiment focuses on the difference from other embodiments, and the various embodiments may be used alone or in combination with each other as needed, for example, a person skilled in the art may combine any two or more method embodiments as needed to implement the information processing method of the present invention, and the method is not limited herein.

< example seven >

The cold start recommendation device provided by the embodiment comprises: the system comprises a user characteristic acquisition module, a product characteristic acquisition module, a matching module and a recommendation module. Wherein:

the user characteristic acquisition module is used for acquiring a user characteristic value of the new user for the set user characteristic according to user data generated by the new user through the third-party application;

the product characteristic acquisition module is used for acquiring a classification characteristic value obtained by classifying each product in the product set according to the user characteristics;

In one embodiment, the user characteristics include at least one of gender characteristics, age characteristics, geographic characteristics, and academic history characteristics.

In an embodiment, the matching module is specifically configured to:

according to the classification characteristic value of each product, the preference degree of the corresponding product to each user characteristic value is obtained;

and obtaining the matching degree between the new user and each product according to the corresponding preference degree of each product and the influence weight of each user characteristic.

In one embodiment, the recommendation module is further configured to:

and acquiring the search heat of each product in the search engine so as to obtain a product recommendation list of the new user at least according to the search heat.

In one embodiment, the recommendation module is further configured to:

and acquiring the transaction popularity of each product formed by transaction payment so as to obtain a product recommendation list of the new user at least according to the transaction popularity.

In one embodiment, the recommendation module further comprises: a comparison unit and a recommendation unit, wherein:

the comparison unit is used for comparing the user characteristic value of the new user with the user characteristic value of the historical new user to obtain a comparison result;

and the recommending unit is used for obtaining the degree of cooperation of each product according to the comparison result and the converted products of the historical new users within the set time after registration, so as to obtain a product recommending list of the new users at least according to the degree of cooperation, wherein the converted products belong to a product set.

In one embodiment, the recommending unit is further configured to obtain a product converted by the historical new user within a set time after registration;

obtaining a classification characteristic value obtained by classifying each converted product in a historical new user according to user characteristics;

In an embodiment, the cold start recommendation device provided in this embodiment further includes a conversion list obtaining module and an updating module, where:

the conversion list acquisition module is used for acquiring a product list actually converted by a new user;

and the updating module is used for comparing the actually converted product list with the product recommendation list and updating the book recommendation list according to the comparison result.

In an embodiment, the cold start recommendation device provided in this embodiment further includes a product set forming module, where: the product set forming module is used for screening out products to be recommended from all the products according to set conditions to form a product set; the set conditions include at least one of a search heat in the search engine exceeding a set search heat, a transaction heat formed through transaction payment exceeding a set transaction heat, and a conversion rate in the historical new user exceeding a set conversion rate.

The product may be a book or the like.

< electronic apparatus >

The electronic device provided by the embodiment comprises any one of the cold start recommending devices in the embodiments.

Alternatively, as shown in fig. 8, an electronic device 80 provided in this embodiment includes: a memory 81 and a processor 82. Wherein,

a memory 81 for storing computer instructions.

And the processor 82 is configured to call the computer instruction from the memory 81 and execute the cold start recommendation method according to any one of the first to fifth embodiments under the control of the computer instruction.

The electronic device may be, for example, the terminal device 1200 in fig. 1, or may be the server 1100 in fig. 1, and may further include the terminal device 1200 and the server 1100 in fig. 1, which is not limited herein.

< storage Medium >

In this embodiment, a computer-readable storage medium is further provided, where the storage medium stores computer instructions, and when the computer instructions in the storage medium are executed by a processor, the method implements any one of the cold start recommendation methods provided in the foregoing embodiments.

The present invention may be a computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions embodied therewith for causing a processor to implement various aspects of the present invention.

The computer readable storage medium may be a tangible device that can hold and store the instructions for use by the instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic memory device, a magnetic memory device, an optical memory device, an electromagnetic memory device, a semiconductor memory device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include a portable computer diskette, a hard disk, a random access memory (RAM, a read-only memory (ROM, an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM, a portable compact disc read-only memory (CD-ROM), a digital versatile disc (DVD, a memory stick, a floppy disk, a mechanical coding device, a punch card or an in-groove protrusion structure having instructions stored thereon, for example, and any suitable combination of the foregoing.

The computer-readable program instructions described herein may be downloaded from a computer-readable storage medium to a respective computing/processing device, or to an external computer or external storage device via a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers. The network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in the respective computing/processing device.

The computer program instructions for carrying out operations of the present invention may be assembler instructions, instruction set architecture (ISA instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages In some embodiments, the various aspects of the present invention are implemented by personalizing electronic circuitry, such as programmable logic circuits, field programmable gate arrays (FPGAs, or Programmable Logic Arrays (PLAs), that execute computer-readable program instructions, with state information of the computer-readable program instructions.

It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer readable program instructions.

These computer-readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer-readable program instructions may also be stored in a computer-readable storage medium that can direct a computer, programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer-readable medium storing the instructions comprises an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.

The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer, other programmable apparatus or other devices implement the functions/acts specified in the flowchart and/or block diagram block or blocks.

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions. It is well known to those skilled in the art that implementation by hardware, by software, and by a combination of software and hardware are equivalent.

Having described embodiments of the present invention, the foregoing description is intended to be exemplary, not exhaustive, and not limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen in order to best explain the principles of the embodiments, the practical application, or improvements made to the technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein. The scope of the invention is defined by the appended claims.

Claims

1. A cold start recommendation method, comprising:

2. The method of claim 1, wherein the user characteristics include at least one of gender characteristics, age characteristics, geographic characteristics, and academic history characteristics.

3. The method of claim 1, wherein the obtaining a degree of matching between the new user and each product according to the user feature value and the classification feature value of each product comprises:

4. The method of claim 1, wherein the method further comprises:

5. The method of claim 1, wherein the method further comprises:

6. The method of claim 1, wherein the method further comprises:

7. The method of claim 6, wherein the obtaining the synergy of each product according to the comparison result and the products converted by the historical new users within the set time after registration comprises:

8. The method of any of claims 1-7, wherein the method, after obtaining the product recommendation list for the new user, further comprises:

obtaining a product list actually converted by the new user;

9. The method of any of claims 1 to 7, wherein the method further comprises:

10. A cold start recommendation method, comprising:

11. The method of claim 10, wherein the method further comprises:

acquiring the search heat of each product in a search engine;

12. A cold start recommendation device comprising:

13. An electronic device comprising the cold start recommendation apparatus of claim 12; alternatively, the electronic device includes: a memory for storing computer instructions and a processor for retrieving said computer instructions from said memory and executing the cold start recommendation method of any one of claims 1-11 under the control of said computer instructions.

14. A computer readable storage medium storing computer instructions which, when executed by a processor, implement a cold start recommendation method as claimed in any one of claims 1-11.