Digital product recharging and selecting system

Abstract

The invention relates to the technical field of digital product data processing, and discloses a digital product recharging and selecting system which comprises a data acquisition module, a feature extraction module, a score calculation module, a score compensation module and a commodity recommendation module. Compared with the prior art that a digital product is subjected to static recommendation only based on historical consumption behaviors of a user, and particularly, real-time dynamic optimization recommendation cannot be realized under the condition of rapid change of user preferences or cold start of new commodities, the method and the device have the advantages that the user behavior feature sensitivity index and the real-time interaction behavior are combined, so that the real-time dynamic optimization recommendation can be realized; according to the method, the weight of the recommendation model is dynamically adjusted, accurate capturing of user requirements and real-time updating of the recommendation sequence are achieved, therefore, the problems that in a traditional recommendation system, recommendation results are homogenized, and the real-time performance is insufficient are solved, and the accuracy of digital product recommendation is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of digital product data processing, and in particular relates to a digital product recharge and selection system. Background Art

[0002] At present, mobile recharge of digital products refers to a way of recharging using mobile phones, tablets or other digital devices through mobile networks or wireless networks. It includes recharging digital products such as mobile phone bills, traffic, and online payments. With the popularization and intelligent development of digital products, users' demand for the convenience, intelligence and accuracy of recharge methods is increasing. However, the existing recharge selection system still has many shortcomings. For example, traditional recommendation systems usually perform static analysis based on users' historical recharge records, lack of full consideration of real-time behavior and contextual features, resulting in insufficient accuracy and personalization of recommendation results. At the same time, for cold start products or low exposure products, it is difficult for existing technologies to balance their recommendation priorities and fully meet users' needs for new products and personalized services. In addition, in user consumption behavior, behavioral characteristics such as price sensitivity and recharge frequency sensitivity have not been effectively utilized, resulting in the recommendation system being unable to dynamically adjust the recommendation strategy according to the actual needs of users. Especially in scenarios where user preferences change rapidly, it is difficult for existing recommendation systems to respond in real time, further reducing user satisfaction and platform conversion rates. Therefore, there is an urgent need for a digital product recharge selection system that can comprehensively consider users' real-time behavior, contextual characteristics and sensitivity indicators, so as to achieve accurate and efficient product recommendations in dynamic and diverse consumption scenarios, thereby improving the personalization, real-time nature and diversity of recommendation effects. Summary of the invention

[0003] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to propose a digital product recharge selection system, which aims to solve the technical problem that in the prior art, digital products are only statically recommended based on the user's historical consumption behavior, especially under the conditions of rapid changes in user preferences or cold start of new products, and real-time dynamic optimization recommendations cannot be achieved.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solution: The present invention provides a digital product recharge selection system, comprising:

[0005] The data collection module is used to collect the user's recharge records, consumption behavior and context data; the recharge records include the recharge amount, time interval and product category, the consumption behavior includes the number of views, click behavior and stay time, and the context data includes the consumption time period and device type;

[0006] The recharge amount and time interval are serialized according to the preset time period, and the recharge amount sequence X and time interval sequence T are obtained through standardization;

[0007] The principal component analysis method is used to reduce the dimension of the collected recharge records, consumption behavior and context data to generate the user behavior feature vector U = {u 1 ,u 2 , ..., u k}, where u 1 ,u 2 ,u k is the user behavior feature component, k is the length of the user behavior feature vector;

[0008] Feature extraction module, used to extract user behavior feature component u from user behavior feature vector U 1 and u 2 , where u 1 is the characteristic component of the average recharge amount, u 2 is the recharge frequency characteristic component;

[0009] Use the recharge amount sequence X and the average recharge amount feature component u 1 Calculate the price sensitivity index S p ;

[0010] Using the time interval sequence T and the recharge frequency characteristic component u 2 Calculate the frequency sensitivity index S f ;

[0011] Score calculation module, used to construct the attribute vector P of each product i i ={p 1 , p 2 , p 3}, where p 1 is the commodity price, p 2 is the popularity of the product, p 3 is the consumption time period weight;

[0012] According to the price sensitivity index S p and frequency sensitivity index S f Dynamically adjust user behavior feature vectors:

[0013] U ′ = {u 1 ·(1+λ p ·S p ),u 2 ·(1+λ f ·S f ), ..., u k}

[0014] Among them, U ′ is the user behavior feature vector after dynamic adjustment, λ p is the price sensitivity weight coefficient, λ fis the frequency sensitivity weight coefficient;

[0015] According to the dynamically adjusted user behavior feature vector U ′ and the attribute vector P of product i i Calculate the matching score M between user and product i;

[0016] The score compensation module is used to obtain the number of clicks on the product in real time and to perform score compensation on the matching score M between the user and the product i. The compensation formula is:

[0017]

[0018] Among them, R is the matching score after compensation, γ is the impact factor of click times, and C i is the number of clicks on product i, and max(C) is the maximum number of clicks among all products;

[0019] The product recommendation module is used to sort the products in descending order according to the compensated matching score R and generate a recommendation sequence.

[0020] Preferably, in the data acquisition module, the standardization processing formula is:

[0021]

[0022] Among them, x 1 is the recharge amount, μ x and δ x is the mean and standard deviation of the recharge amount, t 1 is the time interval, ρ t and σ t is the mean and standard deviation of the time interval.

[0023] Preferably, in the feature extraction module, the price sensitivity index S p The calculation formula is:

[0024]

[0025] Among them, S p is the price sensitivity index, max is the maximum function, and min is the minimum function.

[0026] Preferably, the frequency sensitivity index S f The calculation formula is:

[0027]

[0028] Where Var(T) is the variance of the time interval series.

[0029] Preferably, in the data collection module, the preset time period is dynamically adjusted according to the user's historical recharge records.

[0030] Preferably, in the score calculation module, the calculation formula for the matching score M between the user and the product i is: M = U ′ ·P i .

[0031] Preferably, the product recommendation module also includes re-scoring and recommending in combination with the user's favorite clicks, and the calculation formula is:

[0032] R ′ =R+τ·F

[0033] Among them, τ is the influence factor of favorite click, F is the favorite status, which takes the value of 0 or 1, and R ′ is the matching score after rescoring.

[0034] The beneficial effect of the present invention is that compared with the prior art in which digital products are statically recommended only based on the user's historical consumption behavior, especially under the conditions of rapid changes in user preferences or cold start of new products, the technical problem of being unable to achieve real-time dynamic optimization recommendation is solved. Since the present application dynamically adjusts the recommendation model weight by combining the user behavior feature sensitivity index and real-time interactive behavior, it achieves accurate capture of user needs and real-time update of the recommendation sequence, thereby avoiding the problems of homogeneity of recommendation results and lack of real-time performance in traditional recommendation systems, and improving the accuracy of digital product recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] 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.

[0036] Figure 1 A schematic diagram of the flow chart of the first embodiment of a digital product recharge and selection system provided by the present invention.

[0037] Figure 2 A schematic diagram of a device for a digital product recharge and selection system provided by the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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.

[0039] Embodiment 1: Figure 1 , which is a flow chart of the first embodiment of the digital product recharge and selection system of the present invention, and proposes the first embodiment of the digital product recharge and selection system of the present invention.

[0040] In the first embodiment, the digital product recharge and selection system includes:

[0041] The data collection module is used to collect the user's recharge records, consumption behavior and context data; the recharge records include the recharge amount, time interval and product category, the consumption behavior includes the number of views, click behavior and stay time, and the context data includes the consumption time period and device type;

[0042] The recharge amount and time interval are serialized according to the preset time period, and the recharge amount sequence X and time interval sequence T are obtained through standardization;

[0043] The principal component analysis method is used to reduce the dimension of the collected recharge records, consumption behavior and context data to generate the user behavior feature vector U = {u 1 ,u 2 , ..., u k}, where u 1 ,u 2 ,u k is the user behavior feature component, k is the length of the user behavior feature vector;

[0044] It should be noted that principal component analysis (PCA) constructs a covariance matrix to reduce the feature dimension of the user's recharge records, consumption behavior and context data, retaining the main components and removing redundant information; each component in the user behavior feature vector reflects the user's behavioral preference in a specific dimension;

[0045] It is understandable that in order to better adapt to the consumption habits and behavioral characteristics of different user groups, the length k of the feature vector can be adjusted according to actual needs; compared with traditional statistical methods, standardization processing and dimensionality reduction technology can more accurately extract the key features of user behavior and improve the computational efficiency and accuracy of subsequent recommendation algorithms.

[0046] Feature extraction module, used to extract user behavior feature component u from user behavior feature vector U 1 and u 2 , where u 1 is the characteristic component of the average recharge amount, u 2 is the recharge frequency characteristic component;

[0047] Use the recharge amount sequence X and the average recharge amount feature component u 1 Calculate the price sensitivity index S p ;

[0048] Using the time interval sequence T and the recharge frequency characteristic component u 2 Calculate the frequency sensitivity index S f ;

[0049] It should be noted that the average recharge amount characteristic component represents the average recharge amount level of users in a certain period of time, and is an important indicator for measuring users' consumption capacity; the recharge frequency characteristic component represents the average recharge frequency of users in a certain period of time, and is used to reflect users' consumption behavior habits; the price sensitivity index is used to measure users' sensitivity to commodities of different prices. For example, if the user's recharge amount range is large and the average recharge amount is small, it means that the user is more sensitive to price changes and the index value is higher; the frequency sensitivity index is used to reflect the stability of users' recharge behavior. If the variance of the time interval series is large, it means that users' recharge behavior is unstable and they are more sensitive to recharge frequency.

[0050] It can be understood that the characteristic component u 1 and u 2 It is a key indicator extracted through comprehensive analysis of users' recharge behavior, which can effectively reflect users' consumption capacity and consumption habits; sensitivity index S p and S f The weight distribution of the recommendation system can be adjusted dynamically to provide more targeted recommendation results for price-sensitive or frequency-sensitive users.

[0051] It should be understood that compared with the traditional system that makes recommendations based only on the recharge amount, the present invention can more accurately capture the user's consumption preferences by introducing price and frequency sensitivity indexes; by quantifying user sensitivity characteristics, the system can dynamically adjust the recommendation logic to make the recommendation results more accurate and personalized.

[0052] Score calculation module, used to construct the attribute vector P of each product i i ={p 1 , p 2 , p 3}, where p 1 is the commodity price, p 2 is the popularity of the product, p 3 is the consumption time period weight;

[0053] According to the price sensitivity index S p and frequency sensitivity index S f Dynamically adjust user behavior feature vectors:

[0054] U ′ = {u 1 ·(1+λ p ·S p ),u 2 ·(1+λ f·S f ), ..., u k}

[0055] Among them, U ′ is the user behavior feature vector after dynamic adjustment, λ p is the price sensitivity weight coefficient, λ f is the frequency sensitivity weight coefficient;

[0056] According to the dynamically adjusted user behavior feature vector U ′ and the attribute vector P of product i i Calculate the matching score M between user and product i;

[0057] It should be noted that the commodity attribute vector is constructed based on the commodity price, popularity and time period weights, which can comprehensively reflect the characteristics of the commodity; the price sensitivity index and frequency sensitivity index are used to adjust the weights of the price and frequency related components in the user behavior feature vector respectively, so that the recommendation results are more in line with user preferences; the dynamically adjusted user behavior feature vector U ′ It can more accurately match product attribute vectors and improve the accuracy of recommendation scores.

[0058] It is understandable that when users are more price sensitive, λ p The weight of will increase, and the system will tend to recommend lower-priced products; when the user is more sensitive to frequency, λ f The weight will increase, and the system will pay more attention to the characteristics related to the frequency of commodity consumption; the higher the matching score M, the higher the degree of match between the user and the commodity, and the higher the priority in the recommendation list.

[0059] It should be understood that the process of dynamically adjusting the user behavior feature vector can more flexibly adapt to the dynamic changes of user behavior and preferences compared to traditional static feature matching; the sensitivity weights of different users can be personalized according to the user's historical behavior, thereby improving the personalization ability of the recommendation system; the p in the product attribute vector 1 , p 2 , p 3 It can also be dynamically adjusted according to the actual attributes of the goods to better reflect the characteristics of the goods.

[0060] The score compensation module is used to obtain the number of clicks on the product in real time and to perform score compensation on the matching score M between the user and the product i. The compensation formula is:

[0061]

[0062] Among them, R is the matching score after compensation, γ is the impact factor of click times, and C iis the number of clicks on product i, and max(C) is the maximum number of clicks among all products;

[0063] It should be noted that the number of clicks indicates the number of times a product is clicked in the system. It is dynamic data that reflects the popularity of a product. The higher the number of clicks, the more attention the product receives from users. The maximum number of clicks indicates the highest number of clicks among all products. It is used to normalize the click data of other products to ensure that the proportion of score compensation is reasonable and uniform. The click impact factor is an adjustable parameter used to control the impact of clicks on the final score. When the γ value is larger, the impact of clicks on the compensation score is more significant, which is suitable for strengthening the recommendation of hot-selling products. When the γ value is smaller, the weight of clicks is lower, which is more suitable for diversified recommendation scenarios.

[0064] It is understandable that the rating compensation mechanism can improve the priority of real-time popular products in the recommendation list through dynamic adjustment of the number of clicks; the use of normalization processing can effectively avoid the deviation of the rating results caused by the uneven distribution of clicks; the number of clicks, as real-time dynamic data, can reflect the user's current focus, making the recommendation results more in line with the user's immediate needs.

[0065] It should be understood that the calculation of the compensation score R is a dynamic optimization of the initial matching score, which can comprehensively consider the real-time click performance of the product on the basis of static matching; by adjusting the size of γ, the system can flexibly adjust the recommendation strength for popular products, which is especially effective in cold start scenarios; this mechanism can not only enhance the exposure rate of popular products, but also to a certain extent solve the problem of unpopular products being ignored due to insufficient clicks.

[0066] The product recommendation module is used to sort the products in descending order according to the compensated matching score R and generate a recommendation sequence.

[0067] It should be noted that the compensation score R of a product is a comprehensive score calculated by combining the initial matching score M with dynamic data such as the number of clicks. The compensation scores of all products are arranged in descending order. Products with higher scores have higher priorities in the recommendation sequence. The system generates a recommendation list by arranging the products in descending order of the compensation score R. For products with the same score, further priority adjustments are made based on the popularity or number of clicks of the products. Constraints on category distribution and price distribution can also be introduced into the recommendation sequence to avoid homogeneity of recommended content and improve the diversity of recommendations and user experience.

[0068] It should be understood that the recommendation results are not static, but are optimized through real-time dynamic data such as number of clicks, collection status, etc., so that the recommended content is closer to the user's current needs; the dynamic change of the compensation score R can reflect the popularity of the product, making popular products easier to be seen by users; through the normalization of the number of clicks and the adjustment of the compensation factor, the system can find an appropriate trade-off between balancing the cold start problem of new products and the priority of popular products; the recommendation sequence can be updated in real time according to the user's browsing, clicking or purchasing behavior, ensuring that the recommendation list is more closely matched with the user's interests.

[0069] In addition, the present invention also provides a digital product recharge selection device, please refer to Figure 2, a digital product recharge and selection device includes: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a digital product recharge and selection system in the above-mentioned embodiment 1. A digital product recharge and selection device in the embodiment of the present invention may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. A digital product recharge and selection device is only an example and should not bring any limitations to the functions and scope of use of the embodiments of the present invention. A digital product recharge selection device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of a digital product recharge selection device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow a digital product recharge selection device to communicate wirelessly or wired with other devices to exchange data. Although a digital product recharge selection device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0070] The present invention also provides a computer program product, including a computer program, which implements the steps of the digital product recharge selection system as described above when the computer program is executed by a processor. The computer program product provided by the present invention can solve the technical problem of digital product recharge selection. Compared with the prior art, the beneficial effects of the computer program product provided by the present invention are the same as the beneficial effects of the digital product recharge selection system provided by the above embodiment, which will not be repeated here.

[0071] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the system of the embodiments disclosed in the present invention are executed.

[0072] It should be understood that the various parts disclosed in the present invention can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

Claims

1. A digital product recharge selection system, characterized in that: The system includes: Data collection module, used to collect users' recharge records, consumption behaviors and context data; Recharge records include recharge amount, time interval and product category; consumption behavior includes number of views, click behavior and stay time; context data includes consumption time period and device type; The recharge amount and time interval are serialized according to the preset time period, and the recharge amount sequence X and time interval sequence T are obtained through standardization; The principal component analysis method is used to reduce the dimension of the collected recharge records, consumption behaviors and context data to generate the user behavior feature vector U = {u1, u2, ..., u k }, where u1, u2, u k is the user behavior feature component, k is the length of the user behavior feature vector; A feature extraction module is used to extract user behavior feature components u1 and u2 from the user behavior feature vector U, where u1 is the average recharge amount feature component and u2 is the recharge frequency feature component; Calculate the price sensitivity index S using the recharge amount sequence X and the average recharge amount feature component u1 p ; Calculate the frequency sensitivity index S using the time interval sequence T and the recharge frequency characteristic component u2 f ; Score calculation module, used to construct the attribute vector P of each product i i ={p1, p2, p3}, where p1 is the price of the product, p2 is the popularity of the product, and p3 is the weight of the consumption time period; According to the price sensitivity index S p and frequency sensitivity index S f Dynamically adjust user behavior feature vectors: U ′ ={u1·(1+λ p ·S p ),u2·(1+λ f ·S f ),...,u k } Among them, U ′ is the user behavior feature vector after dynamic adjustment, λ p is the price sensitivity weight coefficient, λ f is the frequency sensitivity weight coefficient; According to the dynamically adjusted user behavior feature vector U ′ and the attribute vector P of product i i Calculate the matching score M between user and product i; The score compensation module is used to obtain the number of clicks on the product in real time and perform score compensation on the matching score M between the user and the product i. The compensation formula is: Among them, R is the matching score after compensation, γ is the impact factor of click times, and C i is the number of clicks on product i, and max(C) is the maximum number of clicks among all products; The product recommendation module is used to sort the products in descending order according to the compensated matching score R and generate a recommendation sequence.

2. A digital product recharge and selection system as claimed in claim 1, characterized in that: Data acquisition module, the standardization processing formula is: Among them, x1 is the recharge amount, μ x and δ x is the mean and standard deviation of the recharge amount, t1 is the time interval, ρ t and σ t is the mean and standard deviation of the time interval.

3. A digital product recharge and selection system as claimed in claim 1, characterized in that: In the feature extraction module, the price sensitivity index S p The calculation formula is: Among them, S p is the price sensitivity index, max is the maximum function, and min is the minimum function.

4. A digital product recharge and selection system as claimed in claim 1, characterized in that: Frequency sensitivity index S f The calculation formula is: Where Var(T) is the variance of the time interval series.

5. A digital product recharge and selection system as claimed in claim 1, characterized in that: In the data collection module, the preset time period is dynamically adjusted according to the user's historical recharge records.

6. A digital product recharge and selection system as claimed in claim 1, characterized in that: In the score calculation module, the calculation formula for the matching score M between user and product i is: M = U ′ ·P i .

7. A digital product recharge and selection system as claimed in claim 1, characterized in that: The product recommendation module also includes re-scoring recommendations based on the user's favorite clicks. The calculation formula is: R ′ =R+τ·F Among them, is the collection click influence factor, is the collection status, with a value of 0 or 1, and is the matching score after re-scoring.

Images