An intelligent information recommendation system for online sales

Abstract

The present invention discloses an intelligent information recommendation system for network sales, which belongs to the technical field of sales information recommendation, and comprises a user module, a retrieval module and a recommendation module; the user module is used for users to input demand information, and send the identified demand information to the retrieval module; the retrieval module is used to retrieve sales information according to the demand information, set target channels, and the target channels refer to various selected e-commerce platforms; the obtained sales information is input into each target channel for retrieval, a large amount of sales information is obtained, the obtained sales information is evaluated, the corresponding recommendation value is obtained, each sales information is arranged in order from high to low according to the recommendation value, and a first sequence is obtained; and corresponding quality characteristic data is supplemented; the recommendation module is used to recommend the sales information to the user, obtain the corresponding first sequence, and recommend the sales information to the user according to the first sequence.

Description

Technical Field

[0001] The invention belongs to the technical field of sales information recommendation, and in particular is an intelligent information recommendation system for network sales. Background Art

[0002] In order to meet the growing shopping and personalized needs of consumers. With the rapid development of the Internet and e-commerce, people can easily shop online, but the numerous commodities and information also bring users the problems of difficulty in selection and information overload.

[0003] Moreover, in actual application, the recommendation effect of the existing sales information recommendation system is not ideal, especially at present, more and more consumers are tired of message push, and basically directly delete the pushed sales information without browsing it, or directly block the message push of various shopping platforms; resulting in poor effect of the existing sales information recommendation system, and for consumers, it also constitutes harassment information, lacks a way to determine relevant sales information, and most consumers cannot understand the details of the sales product information, and have concerns about product quality, being deceived, etc.; therefore, in order to solve the above problems and realize sales information recommendation for consumers, the present invention provides an intelligent information recommendation system for online sales. Summary of the invention

[0004] In order to solve the problems existing in the above solutions, the present invention provides an intelligent information recommendation system for online sales.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] An intelligent information recommendation system for online sales, comprising a user module, a retrieval module and a recommendation module;

[0007] The user module is used for the user to input demand information and send the identified demand information to the search module;

[0008] The retrieval module is used to retrieve sales information based on demand information and set target channels, which refer to various selected e-commerce platforms; input the obtained demand information into each target channel for retrieval, obtain a large amount of sales information, evaluate the obtained sales information, obtain corresponding recommendation values, arrange each sales information in order from high to low according to the recommendation value, and obtain a first sequence; and supplement the corresponding quality characteristic data.

[0009] Further, methods for setting up target channels include:

[0010] Identify the available sales channels and calculate the average recommendation value corresponding to each product category in the sales channel;

[0011] Obtain the market share corresponding to each sales channel, and match the corresponding correction coefficient based on the market share;

[0012] The obtained correction coefficient and average recommended value are marked as ε and PTZ respectively, and the corresponding screening value SP is calculated according to the screening formula SP=ε×PTZ, and the product categories with screening values ​​lower than the threshold X2 are removed from the sales channel;

[0013] Set the target channel based on the processing results of each product category and add the corresponding target category label.

[0014] Furthermore, the average recommended value is calculated by sampling.

[0015] Furthermore, the recommendation value is calculated based on the quality and price of the product corresponding to the sales information.

[0016] Furthermore, the method for evaluating the recommended value includes:

[0017] Identify the quality data and price data corresponding to each sales information, evaluate the obtained quality data, obtain the corresponding quality value, and mark the obtained quality value as PZ; set the corresponding discount value and mark value based on the price data, and mark the obtained discount value and mark value as ZX and BZ respectively;

[0018] Calculate the corresponding recommended value according to the recommended value formula TZ=PZ / (BZ-ZX).

[0019] Furthermore, the recommended value of each sales information is optimized according to the user's usage record, the number of times the user selects each target channel is counted in real time, the usage proportion of each target channel is calculated, and each usage proportion is integrated into an optimization feature. The optimization feature is analyzed to obtain the optimization coefficient corresponding to each target channel, which is marked as c;

[0020] Identify the target channel corresponding to each sales information and match the corresponding optimization coefficient, then the calculation formula of the recommended value is optimized to: TZ=c×PZ / (BZ-ZX).

[0021] Furthermore, the method for evaluating the quality data includes:

[0022] Identify each quality item in the quality data, mark it as i, i=1, 2, ..., n, n is a positive integer; evaluate each quality item data to obtain the corresponding single quality value, mark the obtained single quality value as PZi, and calculate the corresponding quality value PZ according to the quality value evaluation formula PZ= (∑PZi) / n.

[0023] Further, the method of analyzing the price data includes:

[0024] Set up a discount channel, input sales information into the discount channel, obtain corresponding discount data, and obtain corresponding discount value based on the discount data; identify the mark value in the sales information.

[0025] Furthermore, the method for obtaining quality characteristic data includes:

[0026] Acquire quality item data whose single quality value is lower than the threshold value X1, extract cause data in the quality item data, where the cause data is the data that makes the single quality value lower than the threshold value X1; acquire the impact data corresponding to the cause data, and combine and integrate each cause data and impact data into quality feature data.

[0027] The recommendation module is used to recommend sales information to users, obtain the corresponding first sequence, and recommend sales information to users according to the first sequence.

[0028] Furthermore, before recommending sales information to the user according to the first sequence, the recommendation module identifies the large items in the first sequence, obtains the user's location, obtains the large item location information corresponding to the large item sold according to the user's location, and adds the obtained large item location information to the first sequence.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] Through the mutual cooperation between the user module, the search module and the recommendation module, intelligent recommendation of sales information for users is realized, breaking down platform barriers and obtaining appropriate sales information based on user needs; and corresponding quality characteristic data is supplemented in the recommended sales information, so that users can intuitively understand the quality and other information of each product, solving the problem that most consumers cannot understand the details of the sales product information and have concerns about product quality, being deceived, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0032] Figure 1 This is a principle block diagram of the present invention;

[0033] Figure 2 This is a working principle diagram of the retrieval module of the present invention. DETAILED DESCRIPTION

[0034] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] like Figure 1 to Figure 2 As shown, an intelligent information recommendation system for online sales includes a user module, a retrieval module and a recommendation module;

[0036] The user module is used for users to input demand information, including product names, price ranges and other information; the identified demand information is sent to the search module; and the user module has a user information management function for users to manage their personal information.

[0037] The retrieval module is used to retrieve sales information according to demand information, and the specific method includes:

[0038] Set target channels, which refer to various selected e-commerce platforms; input the obtained demand information into each target channel for retrieval to obtain a large amount of sales information, including corresponding source information; evaluate the obtained sales information to obtain the corresponding recommendation value, arrange the sales information in descending order according to the recommendation value to obtain the first sequence; and supplement the corresponding quality characteristic data.

[0039] Ways to set up a goal funnel include:

[0040] Obtain various sales channels currently available on the market, as long as they are regular and safe sales channels that can be retrieved and traded; evaluate the recommended value corresponding to each product information. The specific method of calculating the recommended value is described in the evaluation of the obtained sales information below; calculate the average recommended value corresponding to the product type. If the number of products is too large, a sampling method can be used to calculate the average recommended value; identify the market share of the sales channel, and match the corresponding correction coefficient according to the market share to correct the sales channel with the first share. The correction coefficient is greater than or equal to 1. Specifically, the correction coefficient corresponding to a certain market share is set manually. For shares that are too low or too high, there is no need to set them one by one. If the correction coefficient is lower than a certain share, it is all 1. If it is higher than a certain share, it directly indicates that the standard is met and can be regarded as infinite. ; Only mark the corresponding coefficients for the shares within the interval one by one, set the correction coefficients one by one with the share unit 1 as the span, and then apply the interpolation method for calculation. For example, if the share is 3.8%, use the interpolation method to calculate according to the correction coefficients corresponding to 3% and 4%; reduce the number of correction coefficients to prevent sales channels with a certain share from being directly eliminated; mark the obtained correction coefficient and average recommended value as ε and PTZ respectively, calculate the corresponding screening value according to the screening formula SP=ε×PTZ, and eliminate the commodity categories with screening values ​​lower than the threshold X2 from the sales channel until all commodity categories of the sales channel are evaluated, and mark the sales channels whose commodity categories have not been completely eliminated as target channels, and add the corresponding target category labels. The category labels refer to the types of commodities that have not been eliminated.

[0041] In other embodiments, the target channel may also be specified manually.

[0042] Methods for evaluating the sales information obtained include:

[0043] Identify the quality data and price data corresponding to each sales information. Quality data refers to data related to product quality, including the corresponding ingredient list, material, manufacturer, comments, etc. Price data refers to relevant data including price, discount coupons, cash back, etc. For example, the current channels for purchasing cash back and discount coupons for various products are evaluated, and the cash back and discount coupon channels are preset by the platform. For example, the current various chat group cash back and discount groups are uniformly marked as discount channels. After the corresponding sales link is entered into the discount channel, the discount coupons, cash back and other data in the corresponding relevant price data are obtained; that is, the discount channel is set by the platform itself;

[0044] Evaluate the corresponding marked value and discounted value based on the price data. The marked value refers to the marked price in the sales information, and the discounted value refers to the comprehensive price of the maximum exemption, rebate, etc. that can be used in combination through various discount channels. For example, the total price of discount plus rebate is the discounted value. According to whether various discount channels can be used in combination, the corresponding discounted value can be obtained by using the existing identification and calculation methods. The obtained discounted value and marked value are marked as ZX and BZ respectively.

[0045] Evaluate the obtained quality data to obtain the corresponding quality value, and mark the obtained quality value as PZ; calculate the corresponding recommended value according to the recommended value formula TZ=PZ / (BZ-ZX).

[0046] Methods for evaluating the quality data obtained include:

[0047] The quality data of each corresponding commodity category has fixed quality items, such as the ingredient list, manufacturer, comments and other data generally included in food commodities. The quality data corresponding to each specific commodity category field should include various quality items, which are set by the expert group according to the direction related to the commodity quality of the corresponding commodity field, such as manufacturer information. According to the manufacturer information, the manufacturer's registration information, place of origin, and other production quality-related data of the manufacturer that can be obtained based on the current big data are understood; and when setting the quality items of commodities in each field, an initial value will be preset, that is, when the quality item data cannot be obtained, the initial value will be used as the quality item data and as the subsequent single quality value of the quality item, because in the actual retrieval process, some information may not be available due to commodity reasons, and the initial values ​​of different fields may not be the same, because the quality of the same quality item in different fields has different proportions in the quality evaluation of the commodity. Therefore, when setting each quality item, the corresponding initial value is set synchronously, and the initial value is a percentage value, that is, the value range is [0, 100];

[0048] Identify each quality item in the quality data, mark it as i, i=1, 2, ..., n, n is a positive integer; evaluate each quality item data to obtain the corresponding single quality value, mark the obtained single quality value as PZi, and calculate the corresponding quality value PZ according to the quality value evaluation formula PZ= (∑PZi) / n, i=1, 2, ..., n.

[0049] Evaluate the data of each quality item, that is, establish corresponding evaluation criteria based on the corresponding normative data for evaluation, such as setting corresponding evaluation criteria based on the latest version of food standards, data on the hazards of relevant additives to the human body, etc.; take the ingredient list as an example, first evaluate whether it meets the food standards, if it does not meet the corresponding single quality value, if it meets the full score of 100, do subtraction, that is, identify whether there are elements that are harmful to the human body, and make corresponding reductions according to their content, type, and number of types to obtain the corresponding single quality value. The corresponding evaluation criteria can be set through various current research data, because for each quality item in each industry field, there will be a set of corresponding quality assessment standards, what kind of material The quality is the best and has the corresponding consensus. Therefore, it can be organized into corresponding evaluation criteria, and the corresponding single quality value can be evaluated in combination with the corresponding evaluation criteria; the existing method can be applied to evaluate each quality item; such as establishing a corresponding quality evaluation model based on a CNN network or a DNN network, and manually establishing a corresponding training set for training, the training set includes various quality data of the product set in simulation and the single quality value corresponding to each quality item set; the quality evaluation model after successful training is used for evaluation to obtain the single quality value corresponding to each quality item; because neural network is a prior art in this field, the specific establishment and training process is not described in detail in the present invention.

[0050] Supplement the corresponding quality characteristic data. The quality characteristic data is to extract the reason data of each single quality value in the quality data being lower than the threshold value X1 from the quality item data, obtain its corresponding impact data, and integrate it into quality characteristic data to help users directly understand the defects of the product, such as a harmful additive in the food, the additive content and the corresponding impact.

[0051] The recommendation module is used to recommend sales information to users, obtain the corresponding first sequence, and recommend sales information to users according to the first sequence.

[0052] The recommended information includes corresponding discount information, such as discount channels, etc., and may also have corresponding tutorials to help users save costs.

[0053] Through the mutual cooperation between the user module, the search module and the recommendation module, intelligent recommendation of sales information for users is realized, breaking down platform barriers and obtaining appropriate sales information based on user needs; and corresponding quality characteristic data is supplemented in the recommended sales information, so that users can intuitively understand the quality and other information of each product, solving the problem that most consumers cannot understand the details of the sales product information and have concerns about product quality, being deceived, etc.

[0054] In one embodiment, it is often difficult to understand the real offline situation through image information during online shopping. For large items, some people do not make online purchases due to inconvenience in returning or exchanging them. Therefore, in order to solve this problem, the following method is proposed:

[0055] Identify the large items in the first sequence, and evaluate the large items from the perspectives of volume and price. If the price or volume exceeds a certain value, the large items are directly identified as large items. For the others, the corresponding weights can be preset for comprehensive calculation. Other existing technologies can also be used to set the corresponding evaluation standards for large items.

[0056] The user location is obtained, and based on the user location, large-item location information of corresponding large-item items for sale around the user is identified, such as the locations of stores selling the large-item items; the obtained large-item location information is added to the first sequence.

[0057] In one embodiment, as the number of times a user uses the service increases, the user's preference for one or more target channels can be understood. Therefore, the first sequence can be optimized based on the user's usage records. The specific method is as follows:

[0058] Real-time statistics are made on the number of times users select each target channel, and the usage ratio of each target channel is calculated. The usage ratios are integrated into optimization features. For example, if the usage ratios are 0.5, 0.2, 0.1, 0.1, 0.05, 0.05, 0, 0, 0, respectively, the optimization features are (0.5, 0.2, 0.1, 0.1, 0.05, 0.05, 0, 0, 0). Based on the CNN network or the DNN network, a corresponding optimization analysis model is established, and a corresponding training set is established manually for training. The training set includes the optimization features of various simulation settings and the optimization coefficients of each target channel set accordingly. After successful training, the optimization analysis model is used for analysis to obtain the optimization coefficients corresponding to each target channel, which are marked as c.

[0059] Identify the target channel corresponding to each sales information and match the corresponding optimization coefficient, then the calculation formula of the recommended value is optimized to: TZ=c×PZ / (BZ-ZX).

[0060] The above formulas are all calculated by removing dimensions and taking numerical values. The formula is a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data.

[0061] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An intelligent information recommendation system for online sales, characterized in that: Includes user module, retrieval module and recommendation module; The user module is used for the user to input demand information and send the identified demand information to the search module; The retrieval module is used to retrieve sales information according to demand information, set target channels, and the target channels refer to various selected e-commerce platforms; input the acquired demand information into each target channel for retrieval, obtain a large amount of sales information, evaluate the acquired sales information, obtain corresponding recommendation values, arrange each sales information in descending order according to the recommendation values, obtain a first sequence; and supplement the corresponding quality characteristic data; The recommendation module is used to recommend sales information to the user, obtain the corresponding first sequence, and recommend sales information to the user according to the first sequence; Ways to set up a goal funnel include: Identify the available sales channels and calculate the average recommendation value corresponding to each product category in the sales channel; Obtain the market share corresponding to each sales channel, and match the corresponding correction coefficient based on the market share; The obtained correction coefficient and average recommended value are marked as ε and PTZ respectively, and the corresponding screening value SP is calculated according to the screening formula SP=ε×PTZ, and the product categories with screening values ​​lower than the threshold X2 are removed from the sales channel; Set the target channel based on the processing results of each product category and add the corresponding target category label.

2. The intelligent information recommendation system for online sales according to claim 1, characterized in that: The average recommended value is calculated by sampling.

3. The intelligent information recommendation system for online sales according to claim 1, characterized in that: The recommendation value is calculated based on the quality and price of the product corresponding to the sales information.

4. The intelligent information recommendation system for online sales according to claim 3, characterized in that: Recommended value assessment methods include: Identify the quality data and price data corresponding to each sales information, evaluate the obtained quality data, obtain the corresponding quality value, and mark the obtained quality value as PZ; set the corresponding discount value and mark value based on the price data, and mark the obtained discount value and mark value as ZX and BZ respectively; Calculate the corresponding recommended value according to the recommended value formula TZ=PZ / (BZ-ZX).

5. The intelligent information recommendation system for online sales according to claim 4, characterized in that: Optimize the recommended value of each sales information according to the user's usage record, count the number of times the user selects each target channel in real time, calculate the usage proportion of each target channel, integrate each usage proportion into optimization features, analyze the optimization features, and obtain the optimization coefficient corresponding to each target channel, marked as c; Identify the target channel corresponding to each sales information and match the corresponding optimization coefficient, then the calculation formula of the recommended value is optimized to: TZ=c×PZ / (BZ-ZX).

6. The intelligent information recommendation system for online sales according to claim 4, characterized in that: Methods for evaluating quality data include: Identify each quality item in the quality data, mark it as i, i=1, 2, ..., n, n is a positive integer; evaluate each quality item data to obtain the corresponding single quality value, mark the obtained single quality value as PZi, and calculate the corresponding quality value PZ according to the quality value evaluation formula PZ= (∑PZi) / n.

7. The intelligent information recommendation system for online sales according to claim 4, characterized in that: Methods for analyzing price data include: Set up a discount channel, input sales information into the discount channel, obtain corresponding discount data, and obtain corresponding discount value based on the discount data; identify the mark value in the sales information.

8. The intelligent information recommendation system for online sales according to claim 6, characterized in that: Methods for obtaining quality characteristic data include: Acquire quality item data whose single quality value is lower than the threshold value X1, extract cause data in the quality item data, where the cause data is the data that makes the single quality value lower than the threshold value X1; acquire the impact data corresponding to the cause data, and combine and integrate each cause data and impact data into quality feature data.

9. The intelligent information recommendation system for online sales according to claim 1, characterized in that: Before recommending sales information to the user according to the first sequence, the recommendation module identifies the large items in the first sequence, obtains the user's location, obtains the large item location information of the corresponding large item according to the user's location, and adds the obtained large item location information to the first sequence.

Images