An electronic component product promotion platform and method based on big data

Abstract

The invention discloses an electronic component product promotion platform and method based on big data, and relates to the field of product promotion; the platform comprises: a data integration module, a multidimensional analysis module and a promotion optimization module; the interest parameters of the user are obtained by performing interest analysis through the user's historical purchase records, browsing records and labels and keywords of product advertisements, so as to realize the establishment of accurate user portraits; at the same time, the homogeneity index between the same advertisements and the target advertisement and the correlation index between the different advertisements and the target advertisement are calculated, so as to realize the quantification of the correlation degree between the advertisements; the invention obtains the recommendation value and the recommendation tendency value according to the user's interest parameters and browsing behavior analysis, and judges whether to execute the homogeneous promotion, heterogeneous promotion or mixed promotion strategy based on the recommendation value and performs iteration to form a dynamically changing promotion strategy, so as to make the advertisements received by the user more diversified and personalized, and improve the user's use experience and satisfaction.

Description

Technical Field

[0001] The present invention relates to the technical field of product promotion, and in particular to a big data-based electronic component product promotion platform and method. Background Art

[0002] With the rapid development of Internet technology and information technology, the popularity of e-commerce has gradually become wider. The electronic component promotion platform refers to an online platform that provides marketing and sales services specifically for electronic component manufacturers and suppliers. Compared with traditional offline promotion methods, the electronic component promotion platform can more effectively reduce promotion costs, improve the accuracy of advertising, and achieve a higher return on advertising investment; therefore, the electronic component promotion platform plays an important role, providing electronic component manufacturers and suppliers with a broader market space and more convenient sales channels;

[0003] Currently, the promotion of electronic component products mainly relies on analyzing users' historical purchase records to generate targeted promotion paths. Although this helps provide users with electronic component product recommendations that are more in line with their needs and interests, this promotion method leads to a large number of homogeneous electronic component products recommended to users, which affects the user experience, resulting in poor user stickiness and unsatisfactory promotion results. Summary of the invention

[0004] The main purpose of the present invention is to provide an electronic component product promotion platform and system based on big data to overcome the problems mentioned in the above background technology.

[0005] To achieve the above-mentioned object, according to one aspect of the present invention, there is provided an electronic component product promotion platform based on big data, the platform comprising: a data integration module, a platform database, a multi-dimensional analysis module and a promotion optimization module;

[0006] The data integration module uses crawler technology to obtain user information and sends it to the platform database for storage; the user information includes the user's purchase history and browsing history;

[0007] The platform database stores all product advertisements on the promotion platform;

[0008] The multidimensional analysis module performs interest analysis based on the user's purchase history and browsing history to obtain the user's interest parameters; at the same time, the product advertisements on the promotion platform are associated with analysis to obtain promotion parameters, and the interest parameters and promotion parameters are sent to the promotion optimization module; the interest parameters include the relevant product set and the promotion base value corresponding to each tag, and the promotion base value corresponding to the product advertisement included in each tag; the promotion parameters include the association relationship between each advertisement and the corresponding relationship index; the association relationship is divided into the same type of relationship and the different type of relationship, and the relationship index is divided into the homogeneous index and the association index;

[0009] The promotion optimization module analyzes the interest parameters, promotion parameters and user browsing behavior to generate corresponding promotion strategies. The specific process is as follows:

[0010] When a user enters the promotion platform through the user terminal, the user's interest parameters are retrieved, and the promotion base value corresponding to each label is multiplied by the set quantity conversion coefficient to obtain the number of advertisements corresponding to each label. The promotion base values ​​corresponding to the product advertisements of each label are selected in descending order until the number of selected product advertisements is consistent with the number of advertisements corresponding to the label; the promotion base value of the label to which the selected product advertisement belongs is retrieved and recorded as WL, and the promotion base value LM corresponding to the product advertisement, and the set formula is used. Calculate and obtain the recommendation value Wd, where d1 and d2 are respectively the set proportional coefficients; generate an initial promotion path for the selected product advertisements in descending order of their corresponding recommendation values, record the selected product advertisements as initial recommended product advertisements, and promote the initial recommended product advertisements to the user end according to the initial promotion path;

[0011] The browsing behavior of users browsing the initial recommended product advertisements is retrieved and analyzed in detail to obtain the recommendation tendency value; when the recommendation tendency value is greater than the maximum value in the set tendency interval, a homogeneous promotion strategy is executed; when the recommendation tendency is within the set tendency interval, a mixed promotion strategy is executed; when the recommendation tendency value is less than the minimum value in the set tendency interval, a heterogeneous promotion strategy is executed.

[0012] Furthermore, the specific process of conducting detailed analysis on the browsing behavior of users browsing the initial recommended product advertisements is as follows:

[0013] Compare and analyze the browsing time with the set time interval to divide the browsing time into primary browsing time, secondary browsing time and tertiary browsing time, and record the initial recommended product advertisements corresponding to the primary browsing time, secondary browsing time and tertiary browsing time as primary initial recommended advertisements, secondary initial recommended advertisements and tertiary initial recommended advertisements respectively; count the number of primary initial recommended advertisements, secondary initial recommended advertisements and tertiary initial recommended advertisements respectively, and record them as B1, B2 and B3 respectively; sum the primary browsing time, secondary browsing time and tertiary browsing time to obtain the primary browsing value, secondary browsing value and tertiary browsing value respectively, and record them as B4, B5 and B6 respectively;

[0014] Substitute B1, B2, B3, B4, B5 and B6 into the set formula The recommended tendency value Bd is calculated, where d3, d4, and d5 are respectively set proportional coefficients, and d3>d4>d5>0.

[0015] Furthermore, the specific process of the homogeneous promotion strategy is:

[0016] Take any initial recommended product advertisement as the target advertisement, retrieve similar advertisements of the target advertisement and their corresponding homogeneity indexes, select the similar advertisement with the largest homogeneity index as the homogeneous first-level iterative continuation advertisement of the target advertisement, select the similar advertisement with the second highest homogeneity index in descending order as the homogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the homogeneous m-level iterative continuation advertisements of each initial recommended advertisement, where m is a positive integer;

[0017] When a homogeneous promotion strategy is generated for the mth time, the homogeneous m-level iterative continuation advertisements corresponding to each initial recommended product advertisement are arranged in the order of each initial recommended product advertisement in the initial recommendation path to form a homogeneous m-level iterative recommendation path, and sent to the user end; the browsing behavior of the user end browsing the homogeneous m-level iterative recommendation path is re-analyzed for promotion.

[0018] Furthermore, the specific steps of the heterogeneous promotion strategy are:

[0019] Take any initial recommended product advertisement as the target advertisement, retrieve the non-same advertisements of the target advertisement and their corresponding correlation indexes, select the non-same advertisement with the smallest correlation index as the heterogeneous first-level iterative continuation advertisement of the target advertisement, select the non-same advertisement with the second highest correlation index in the order of small to large as the heterogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the heterogeneous m-level iterative continuation advertisements of each initial recommended advertisement.

[0020] When a heterogeneous promotion strategy is generated for the mth time, the heterogeneous m-level iterative continuation advertisements corresponding to each initial recommended product advertisement are arranged in the order of each initial recommended product advertisement in the initial recommendation path to form a heterogeneous m-level iterative recommendation path, and sent to the user end; the browsing lines of the user end browsing the homogeneous m-level iterative recommendation path are re-analyzed for promotion.

[0021] Furthermore, the specific steps of the mixed promotion strategy are:

[0022] Randomly select half of the initial recommended advertisements in the initial recommended path as target advertisements for homogeneous promotion, and the other half as target advertisements for heterogeneous promotion; retrieve the same type of advertisements as target advertisements for homogeneous promotion and their corresponding homogeneity indexes, select the same type of advertisement with the largest homogeneity index as the homogeneous first-level iterative continuation advertisement of its target advertisement, select the same type of advertisement with the second highest homogeneity index in descending order as the homogeneous second-level iterative continuation advertisement of its target advertisement, and so on to obtain the homogeneous m-level iterative continuation advertisements of each initial recommended advertisement;

[0023] When m times of generated mixed promotion strategies are received, the homogeneous m-level iterative continuation advertisements or heterogeneous m-level iterative continuation advertisements that are their target advertisements are arranged in the order of their corresponding target advertisements in the initial recommendation path to form a mixed m-level iterative recommendation path, the mixed m-level iterative recommendation path is sent to the user end, and the browsing behavior of the user end browsing the mixed m-level iterative recommendation path is re-analyzed for promotion.

[0024] Furthermore, the specific steps of interest analysis are:

[0025] Traverse the product advertisements on the platform and classify them according to their product types, and use the product types as labels, thereby obtaining labels for each product promotion; use TF-IDF technology to extract several keywords corresponding to each product advertisement, so that each product advertisement has a classification label and several keywords;

[0026] Calculate the time difference between the purchase time of each purchase and the current system time to obtain the relevant duration of each purchase, and accumulate the relevant duration to the corresponding purchased product;

[0027] The time difference between the start time and the end time of each browsing is calculated to obtain the browsing duration of each browsing, which is recorded as G1. The time difference between the end time of each browsing and the current time of the system is calculated to obtain the relevant duration corresponding to each browsing number, which is recorded as G2. The browsing duration G1 and the relevant duration G2 are substituted into the set formula. Calculate and obtain the browsing related value Ga corresponding to each browsing number, where a1 and a2 are respectively the set proportional coefficients; thereby accumulating each browsing related value to the corresponding browsing product;

[0028] The products that the user has purchased and browsed are organized into a set of related products for the user, where each product is recorded as a related product for the user; thus, the cumulative number of purchase-related durations of each related product and the corresponding related duration for each accumulation are recorded as Mi, the cumulative number of browsing-related values ​​and the corresponding browsing-related value for each accumulation are recorded as Lj; where i=1, 2, 3…I, j=1, 2, 3…J, I and J are both positive integers, I represents the total cumulative number of purchase-related durations, J represents the total cumulative number of browsing-related values, i represents the cumulative number of any purchase-related durations, j represents the cumulative number of any browsing-related values; Mi represents the purchase-related duration corresponding to the cumulative number of any purchase-related duration, and Lj represents the browsing-related value corresponding to the cumulative number of any browsing-related values; using the set formula Calculate and obtain the base promotion value LM of each product advertisement, where a3 and a4 are respectively the set proportional coefficients; retrieve the label of each product advertisement, and classify each product advertisement according to the label, so that each label includes several product advertisements, and the sum of the promotion base values ​​of the product advertisements in the same label is calculated as the base promotion value of the label.

[0029] Furthermore, the specific steps of association analysis are:

[0030] Retrieve all product advertisements in the promotion platform, select one of the product advertisements as the target advertisement, and use the other product advertisements as reference advertisements. When the reference advertisement and the target advertisement are in the same label, the reference advertisement is recorded as a similar advertisement; when the reference advertisement and the target advertisement do not belong to the same label, the reference advertisement is recorded as a different advertisement; count the intersection keywords and union keywords between similar advertisements and target advertisements, and use TF-IDF technology to extract the TF-IDF scores of each intersection keyword in similar advertisements and target advertisements. It can be seen that one intersection keyword corresponds to two TF-IDF scores, namely the TF-IDF score of the intersection keyword in the similar advertisement and the TF-IDF score of the intersection keyword in the target advertisement; sum up the two TF-IDF scores corresponding to the intersection keyword to obtain the intersection score of the intersection keyword, which is recorded as TFk, where k=1, 2, 3...K, K is a positive integer, K represents the total number of intersection keywords between similar advertisements and target advertisements, and k represents the sequence number of any intersection keyword;

[0031] Compare and analyze the intersection score with the set score interval to divide the intersection score into a high intersection score, a medium intersection score and a low intersection score, count the number of high intersection scores, medium intersection scores and low intersection scores respectively, and record them as S1, S2 and S3 respectively; count the number of intersection keywords and union keywords, and record them as S4 and S5 respectively;

[0032] Substitute S1, S2, S3, S4, S5 and TFk into the given formula Calculate and obtain the homogeneity index Sb between the same type of advertisement and the target advertisement, where b1, b2, and b3 are respectively the set proportional coefficients, and b1>b2>b3>0, and λ1 is the set quantitative coefficient corresponding to the same type of advertisement;

[0033] Retrieve the target advertisement and the products represented by the non-identical advertisement, and determine the relationship between the two products, where the relationship is divided into: direct connection relationship, indirect connection relationship, and the order of indirect connection is recorded as n; set different connection relationships to correspond to a connection coefficient recorded as A1; count the number of intersection keywords and union keywords between the target advertisement and the non-identical advertisement, and record them as A2 and A3 respectively; through the set formula group The correlation index Ab between the non-identical advertisements and the target advertisement is calculated, wherein b4, b5, b6, and b7 are respectively set proportional coefficients; λ2 is the set quantitative coefficient corresponding to the non-identical advertisements, and λ1>λ2>0.

[0034] In order to achieve the above object, according to another aspect of the present invention, a method for promoting electronic components products based on big data is provided, the method comprising the following steps:

[0035] 801: Obtain user information and send it to the platform database for storage; the platform database stores all product advertisements on the promotion platform;

[0036] 802: performing interest analysis based on the user's purchase history and browsing history to obtain interest parameters of the user; performing correlation analysis on the product advertisements of the promotion platform to obtain promotion parameters, and sending the interest parameters and promotion parameters to the promotion optimization module; wherein the interest parameters include the relevant product set and the promotion base value corresponding to each tag therein, and the promotion base value corresponding to the product advertisement included in each tag; the promotion parameters include the correlation relationship between each advertisement and the corresponding relationship index; the correlation relationship is divided into the same type relationship and the different type relationship, and the relationship index is divided into the homogeneous index and the correlation index;

[0037] 803: When the user enters the promotion platform through the user terminal, the user's interest parameter is retrieved, the promotion base value corresponding to each tag is multiplied by the set quantity conversion coefficient to obtain the number of advertisements corresponding to each tag, and the promotion base values ​​corresponding to the product advertisements of each tag are selected in descending order until the number of selected product advertisements is consistent with the number of advertisements corresponding to the tag; the promotion base value of the tag to which the selected product advertisement belongs and the promotion base value corresponding to the product advertisement are retrieved, and a formula calculation and analysis are performed to obtain a recommended value; the selected product advertisement is generated into an initial promotion path in descending order according to its corresponding recommended value, and the selected product advertisement is recorded as an initial recommended product advertisement, and the initial recommended product advertisement is promoted to the user terminal according to the initial promotion path;

[0038] 804: retrieve the browsing behavior of the user browsing the initial recommended product advertisement, and perform detailed analysis on it to obtain the recommendation tendency value; when the recommendation tendency value is greater than the maximum value in the set tendency interval, execute the homogeneous promotion strategy of 805; when the recommendation tendency is within the set tendency interval, execute the mixed promotion strategy of 807; when the recommendation tendency value is less than the minimum value in the set tendency interval, execute the heterogeneous promotion strategy of 806;

[0039] 805: Homogeneous promotion strategy:

[0040] Take any initial recommended product advertisement as the target advertisement, retrieve similar advertisements of the target advertisement and their corresponding homogeneity indexes, select the similar advertisement with the largest homogeneity index as the homogeneous first-level iterative continuation advertisement of the target advertisement, select the similar advertisement with the second highest homogeneity index in descending order as the homogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the homogeneous m-level iterative continuation advertisements of each initial recommended advertisement, where m is a positive integer;

[0041] When a homogeneous promotion strategy is generated for the mth time, the homogeneous m-level iterative continuation advertisements corresponding to the initial recommended product advertisements are arranged in the order of the initial recommended product advertisements in the initial recommendation path to form a homogeneous m-level iterative recommendation path, and the homogeneous m-level iterative recommendation path is sent to the user terminal; the browsing behavior of the user terminal browsing the homogeneous m-level iterative recommendation path is updated to 801;

[0042] 806: Heterogeneous promotion strategy:

[0043] Take any initial recommended product advertisement as the target advertisement, retrieve the non-same advertisements of the target advertisement and their corresponding correlation indexes, select the non-same advertisement with the smallest correlation index as the heterogeneous first-level iterative continuation advertisement of the target advertisement, select the non-same advertisement with the second highest correlation index in the order of small to large as the heterogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the heterogeneous m-level iterative continuation advertisements of each initial recommended advertisement.

[0044] When the heterogeneous promotion strategy is generated for the mth time, the heterogeneous m-level iterative continuation advertisements corresponding to the initial recommended product advertisements are arranged according to the order of the initial recommended product advertisements in the initial recommendation path to form a heterogeneous m-level iterative recommendation path, and sent to the user end; the browsing line of the user end browsing the homogeneous m-level iterative recommendation path is updated to 801;

[0045] 807: Mixed promotion strategy:

[0046] Randomly select half of the initial recommended advertisements in the initial recommended path as target advertisements for homogeneous promotion, and the other half as target advertisements for heterogeneous promotion; retrieve the same type of advertisements as target advertisements for homogeneous promotion and their corresponding homogeneity indexes, select the same type of advertisement with the largest homogeneity index as the homogeneous first-level iterative continuation advertisement of its target advertisement, select the same type of advertisement with the second highest homogeneity index in descending order as the homogeneous second-level iterative continuation advertisement of its target advertisement, and so on to obtain the homogeneous m-level iterative continuation advertisements of each initial recommended advertisement;

[0047] When m times of generating a hybrid promotion strategy are received, the homogeneous m-level iterative continuation advertisements or heterogeneous m-level iterative continuation advertisements as their target advertisements are arranged in the order of their corresponding target advertisements in the initial recommendation path to form a hybrid m-level iterative recommendation path, the hybrid m-level iterative recommendation path is sent to the user end, and the browsing behavior of the user end browsing the hybrid m-level iterative recommendation path is updated to 801.

[0048] Beneficial effects of the present invention:

[0049] The present invention obtains user interest parameters through interest analysis based on the user's historical purchase records, browsing records, and product advertisement tags and keywords, thereby achieving accurate user portrait establishment; at the same time, for the association relationship between different product advertisements, by calculating the homogeneity index between the same type of advertisements and the target advertisement and the association index between different advertisements and the target advertisement, the degree of association between advertisements is quantified, which is conducive to identifying the similarities and differences between advertisements; and provides technical support for more accurate and personalized advertisement recommendation services;

[0050] The present invention obtains recommendation values ​​and recommendation tendency values ​​according to the user's interest parameters and browsing behavior analysis, and then determines whether to implement homogeneous promotion, heterogeneous promotion or mixed promotion strategies, and iterates to form a dynamically changing promotion strategy, so that the advertisements received by users are more diversified and personalized, and the user's usage experience and satisfaction are improved; personalized and accurate electronic component product promotion is achieved, and user satisfaction is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the accompanying drawings:

[0052] Figure 1 It is a schematic diagram of a platform of the present invention. DETAILED DESCRIPTION

[0053] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0054] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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 should fall within the scope of protection of the present invention.

[0055] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so as to describe the embodiments of the present invention described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0056] It should be noted that the relevant information (including but not limited to user information) and data (including but not limited to data for display, data for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties.

[0057] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0058] According to an embodiment of the present invention, Figure 1 As shown, a big data-based electronic component product promotion platform is provided, which includes: a data integration module, a platform database, a multi-dimensional analysis module and a promotion optimization module;

[0059] The data integration module uses crawler technology to obtain user information and sends it to the platform database for storage; the user information includes the user's purchase history and browsing history;

[0060] The platform database stores all product advertisements on the promotion platform;

[0061] The multi-dimensional analysis module extracts promotion parameters by analyzing multi-dimensional information, so as to achieve in-depth analysis of user behavior and lay the foundation for appropriate personalized promotion; specifically:

[0062] Interest analysis:

[0063] Traverse the product advertisements on the platform and classify them according to their product types, and use the product types as labels, thereby obtaining labels for each product promotion; use TF-IDF (term frequency-inverse document frequency) technology to extract several keywords corresponding to each product advertisement, so that each product advertisement has a classification label and several keywords to facilitate product advertisement positioning, improve the searchability and recommendation effect of advertisements, and help provide more accurate and personalized advertising services;

[0064] Retrieve the user's historical purchase records and browsing records; the historical purchase records include the number of historical purchases, the purchase time of each purchase, and the purchased products; the browsing records include the search terms, the number of views, and the start and end time of each browse;

[0065] The time difference between the purchase time of each purchase and the current time of the system is calculated to obtain the relevant duration of each purchase. It should be noted that the larger the relevant duration, the smaller the purchase behavior corresponding to the relevant duration has an impact on the user's current interest preference; conversely, the smaller the relevant duration, the greater the impact of the purchase behavior corresponding to the relevant duration on the user's current interest preference; thus, the relevant duration is accumulated to the corresponding purchased product;

[0066] The time difference between the start time and the end time of each browsing is calculated to get the browsing duration of each browsing, which is recorded as G1. The time difference between the end time of each browsing and the current time of the system is calculated to get the relevant duration corresponding to each browsing number, which is recorded as G2. Using the set formula Calculate and obtain the browsing related value Ga corresponding to each browsing number, where a1 and a2 are respectively the set proportional coefficients; thereby accumulating each browsing related value to the corresponding browsing product;

[0067] The products that the user has purchased and browsed are organized into a set of related products for the user, where each product is recorded as a related product for the user; thus, the cumulative number of purchase-related durations of each related product and the corresponding related duration for each accumulation are recorded as Mi, the cumulative number of browsing-related values ​​and the corresponding browsing-related value for each accumulation are recorded as Lj; where i=1, 2, 3…I, j=1, 2, 3…J, I and J are both positive integers, I represents the total cumulative number of purchase-related durations, J represents the total cumulative number of browsing-related values, i represents the cumulative number of any purchase-related durations, j represents the cumulative number of any browsing-related values; Mi represents the purchase-related duration corresponding to the cumulative number of any purchase-related duration, and Lj represents the browsing-related value corresponding to the cumulative number of any browsing-related values; using the set formula Calculate and obtain the promotion base value LM of each product advertisement, where a3 and a4 are respectively the set proportional coefficients; retrieve the label of each product advertisement, classify each product advertisement according to the label, so that each label includes several product advertisements, and sum up the promotion base values ​​of the product advertisements in the same label to obtain the sum value as the promotion base value of the label;

[0068] The relevant product set, the promotion base value corresponding to each tag therein, and the promotion base value corresponding to the product advertisement included in each tag are used as the user's interest parameter;

[0069] Product Advertising Analysis:

[0070] Retrieve all product advertisements in the promotion platform, select one of the product advertisements as the target advertisement, and the other product advertisements as reference advertisements. When the reference advertisement and the target advertisement are in the same label, the reference advertisement is recorded as a similar advertisement; when the reference advertisement and the target advertisement are not in the same label, the reference advertisement is recorded as a different advertisement;

[0071] Count the intersection keywords and union keywords between the same type of advertisements and the target advertisement, where the intersection keywords refer to the same keywords between the same type of advertisements and the target advertisement, and the union keywords refer to all the keywords contained in both the same type of advertisements and the target advertisement; use the TF-IDF (term frequency-inverse document frequency) technology to obtain the TF-IDF scores of each intersection keyword in the same type of advertisements and the target advertisement, and record them as Tk and Fk respectively, where k=1, 2, 3...K, K is a positive integer, K represents the total number of intersection keywords between the same type of advertisements and the target advertisement, and k represents the serial number of any intersection keyword; it can be seen that the intersection keyword with serial number k corresponds to two TF-IDF scores, namely Tk and Fk, where Tk is the TF-IDF score of the intersection keyword with serial number k in the same type of advertisement, Fk is the TF-IDF score of the intersection keyword with serial number k in the target advertisement; the two TF-IDF scores corresponding to the intersection keywords are summed up to obtain the intersection score of the intersection keywords, recorded as TFk; the intersection score is compared and analyzed with the set score interval. When the intersection score is greater than the maximum value in the set score interval, the intersection score is recorded as a high intersection score; when the intersection score is within the set score interval, the intersection score is recorded as a moderate intersection score; when the intersection score is less than the minimum value in the set score interval, the intersection score is recorded as a low intersection score. The number of high intersection scores, moderate intersection scores and low intersection scores are counted respectively, and they are recorded as S1, S2 and S3 respectively; the number of intersection keywords and union keywords are counted, and they are recorded as S4 and S5 respectively; using the set formula The homogeneity index Sb between the same type of advertisement and the target advertisement is calculated, where b1, b2, and b3 are the set proportional coefficients, and b1>b2>b3>0, and λ1 is the set quantitative coefficient corresponding to the same type of advertisement. It can be seen from the formula that when the intersection keywords are larger, the closer they are to the number of union keywords (i.e. The closer it is to 0), the greater the overlap between the same type of advertisements and the target advertisement, that is, the greater the homogeneity index; thus, the homogeneity index between each type of advertisement and the target advertisement can be obtained;

[0072] It should be noted that TF (Term Frequency) refers to the frequency of a word appearing in a document. The more times a word appears in a document, the greater the TF value, which indicates the importance of the word in the current document. IDF (Inverse Document Frequency) refers to the logarithm of the inverse number of documents containing a word. If a word appears in many documents, its IDF value will be smaller. On the contrary, if a word appears in only a few documents, its IDF value will be larger. IDF can measure the importance of a word to the entire document collection. The TF-IDF value is the product of TF and IDF, which can be used to measure the importance of a word in a specific document and its importance relative to the entire document collection.

[0073] Retrieve the target advertisement and the product represented by the non-identical advertisement, and determine the relationship between the two products, wherein the relationship is divided into: direct connection relationship, indirect connection relationship, and the order of indirect connection is recorded as n; direct connection relationship means that the products represented by the target advertisement and the non-identical advertisement are usually directly connected when used, indirect connection relationship means that the products represented by the target advertisement and the non-identical advertisement are usually indirectly connected when used, and n-order means that the products represented by the target advertisement and the non-identical advertisement are usually separated by n other products when used, and n is a positive integer; when n=1, the first order means that the products represented by the target advertisement and the non-identical advertisement are usually separated by one other product when used; when n=2, the second order means that the products represented by the target advertisement and the non-identical advertisement are usually separated by two other products when used;

[0074] Different connection relationships are set to correspond to a connection coefficient denoted as A1, and the connection coefficient corresponding to the direct connection relationship is greater than the connection coefficient corresponding to the indirect connection relationship;

[0075] Count the number of intersection keywords and union keywords between the target advertisement and non-same advertisements, and record them as A2 and A3 respectively; use the set formula group The correlation index Ab between the non-identical advertisements and the target advertisement is calculated, where b4, b5, b6, and b7 are respectively the set proportional coefficients; λ2 is the set quantitative coefficient corresponding to the non-identical advertisements, and λ1>λ2>0; it can be seen from the formula that when the number of intersection keywords between the non-identical advertisements and the target advertisements is more and the number of union keywords is less, the correlation index is greater; when the order of the connection relationship between the non-identical advertisements and the products represented by the target advertisements is greater, the correlation index is smaller; thus, the correlation index between each non-identical advertisement and the target advertisement can be obtained;

[0076] The association relationship between each advertisement (where the association relationship refers to the same type of relationship and the different type of relationship, the same type of relationship means that the two advertisements belong to the same label, and the different type of relationship means that the two advertisements belong to different labels), and the corresponding relationship index (where the relationship indexes are the homogeneity index and the association index, the homogeneity index refers to the relationship index between two advertisements in the same type of relationship, and the association index refers to the relationship index between two advertisements in the different type of relationship) are recorded as promotion parameters;

[0077] Send interest parameters and promotion parameters to the optimization push module;

[0078] Through the interest analysis function, interest analysis is performed based on the user's historical purchase records, browsing records, and product advertisement tags and keywords to obtain the user's interest parameters, thereby achieving the establishment of an accurate user portrait. At the same time, in view of the correlation between different product advertisements, the homogeneity index between similar advertisements and target advertisements and the correlation index between different advertisements and target advertisements are calculated to quantify the degree of correlation between advertisements, which is conducive to identifying similarities and differences between advertisements; providing technical support for more accurate and personalized advertising recommendation services.

[0079] The promotion optimization module executes corresponding promotion strategies based on the user's current browsing behavior, realizes dynamic changes in homogeneous promotion, mixed promotion and differentiated promotion, so as to meet the needs of users and realize long-chain, flexible and accurate push of electronic component products; specifically:

[0080] Step 1: When the user enters the promotion platform through the user terminal, the user's interest parameters are retrieved, where the interest parameters include the relevant product set and the promotion base value corresponding to each label, and the promotion base value corresponding to the product advertisement included in each label; the promotion base value corresponding to each label is multiplied by the set quantity conversion coefficient to obtain the number of advertisements corresponding to each label, and the corresponding number of product advertisements are selected according to the number of advertisements corresponding to each label, where the promotion base values ​​corresponding to the product advertisements of each label are selected in descending order until the number of selected product advertisements is consistent with the number of advertisements corresponding to the label; the promotion base value of the label to which the selected product advertisement belongs is retrieved and recorded as WL, and the promotion base value LM corresponding to the product advertisement, and the set formula is used. Calculate and obtain the recommendation value Wd, where d1 and d2 are respectively the set proportional coefficients; generate an initial promotion path for the selected product advertisements in descending order of their corresponding recommendation values, record the selected product advertisements as initial recommended product advertisements, and promote the initial recommended product advertisements to the user end according to the initial promotion path;

[0081] Step 2: retrieve the user's browser behavior of browsing the initial recommended product advertisement, where the browsing behavior includes the browsing time of each initial recommended product advertisement, compare and analyze the browsing time with the set time interval, and when the browsing time is greater than the set time interval, the browsing time is recorded as the first-level browsing time, and the initial recommended product advertisement corresponding to the browsing time is recorded as the first-level initial recommended advertisement; when the browsing time is within the set time interval, the browsing time is recorded as the second-level browsing time, and the initial recommended product advertisement corresponding to the browsing time is recorded as the second-level Initial recommended advertisement; when the browsing time is less than the set time interval, the browsing time is recorded as the third-level browsing time, and the initial recommended product advertisement corresponding to the browsing time is recorded as the third-level initial recommended advertisement; the number of first-level initial recommended advertisements, second-level browsing time and third-level browsing time are counted respectively, and recorded as B1, B2 and B3 respectively; the first-level browsing time, second-level browsing time and third-level browsing are summed up to obtain the first-level browsing value, second-level browsing value and third-level browsing value, and recorded as B4, B5 and B6 respectively; using the set formula Calculate and obtain the recommendation tendency value Bd, where d3, d4, and d5 are respectively the set proportional coefficients, and d3>d4>d5>0; when the recommendation tendency value is greater than the maximum value in the set tendency interval, it means that the user is at the stage of being interested in homogeneous products, and the homogeneous promotion strategy of step three is executed; when the recommendation tendency is within the set tendency interval, the mixed promotion strategy of step five is executed; when the recommendation tendency value is less than the minimum value in the set tendency interval, the heterogeneous promotion strategy of step four is executed;

[0082] Step 3: The specific process of homogeneous promotion strategy is:

[0083] Take any initial recommended product advertisement as the target advertisement, retrieve similar advertisements of the target advertisement and their corresponding homogeneity indexes, select the similar advertisement with the largest homogeneity index as the homogeneous first-level iterative continuation advertisement of the target advertisement, select the similar advertisement with the second highest homogeneity index in descending order as the homogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the homogeneous m-level iterative continuation advertisements of each initial recommended advertisement, where m is a positive integer;

[0084] Thus, the homogeneous first-level iterative continuation advertisements corresponding to each initial recommended product advertisement are obtained, and the homogeneous first-level iterative continuation advertisements are arranged in the order of their corresponding initial recommended product advertisements in the initial recommendation path to form a homogeneous first-level iterative recommendation path, and the homogeneous first-level iterative recommendation path is sent to the user terminal, and the browsing behavior of the user terminal browsing the homogeneous first-level iterative recommendation path is sent to step 1;

[0085] When a homogeneous promotion strategy is generated for the second time, the homogeneous secondary iterative continuation advertisements corresponding to each initial recommended product advertisement are arranged in the order of each initial recommended product advertisement in the initial recommendation path to form a homogeneous secondary iterative recommendation path, and the path is sent to the user end; by analogy, when a homogeneous promotion strategy is generated for the mth time, the homogeneous m-level iterative continuation advertisements corresponding to each initial recommended product advertisement are arranged in the order of each initial recommended product advertisement in the initial recommendation path to form a homogeneous m-level iterative recommendation path, and the path is sent to the user end, and the browsing behavior of the user end browsing the homogeneous m-level iterative recommendation path is sent to step one;

[0086] Step 4: The specific process of heterogeneous promotion strategy is:

[0087] Take any initial recommended product advertisement as the target advertisement, retrieve the non-same advertisements of the target advertisement and their corresponding correlation indexes, select the non-same advertisement with the smallest correlation index as the heterogeneous first-level iterative continuation advertisement of the target advertisement, select the non-same advertisement with the second highest correlation index in the order of small to large as the heterogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the heterogeneous m-level iterative continuation advertisements of each initial recommended advertisement.

[0088] Thus, the heterogeneous first-level iterative continuation advertisements corresponding to each initial recommended product advertisement are obtained, and the heterogeneous first-level iterative continuation advertisements are arranged in the order of their corresponding initial recommended product advertisements in the initial recommendation path to form a heterogeneous first-level iterative recommendation path, and the heterogeneous first-level iterative recommendation path is sent to the user end, and the browsing behavior of the user end browsing the heterogeneous first-level iterative recommendation path is sent to step 1;

[0089] When the heterogeneous promotion strategy is generated for the second time, the heterogeneous secondary iterative continuation advertisements corresponding to each initial recommended product advertisement are arranged in the order of each initial recommended product advertisement in the initial recommendation path to form a heterogeneous secondary iterative recommendation path, and the path is sent to the user end; by analogy, when the heterogeneous promotion strategy is generated for the mth time, the heterogeneous m-level iterative continuation advertisements corresponding to each initial recommended product advertisement are arranged in the order of each initial recommended product advertisement in the initial recommendation path to form a heterogeneous m-level iterative recommendation path, and the path is sent to the user end, and the browsing behavior of the user end browsing the homogeneous m-level iterative recommendation path is sent to step one;

[0090] Step 5: The specific process of the mixed promotion strategy is:

[0091] Randomly select half of the initial recommended advertisements in the initial recommended path as target advertisements for homogeneous promotion, and the other half as target advertisements for heterogeneous promotion; retrieve the same type of advertisements as target advertisements for homogeneous promotion and their corresponding homogeneity indexes, select the same type of advertisement with the largest homogeneity index as the homogeneous first-level iterative continuation advertisement of its target advertisement, select the same type of advertisement with the second highest homogeneity index in descending order as the homogeneous second-level iterative continuation advertisement of its target advertisement, and so on to obtain the homogeneous m-level iterative continuation advertisements of each initial recommended advertisement;

[0092] Retrieve the non-identical advertisements as the target advertisements for the heterogeneous promotion and their corresponding correlation indexes, select the non-identical advertisement with the smallest correlation index as the heterogeneous first-level iterative continuation advertisement of the target advertisement, select the non-identical advertisement with the second highest correlation index in the order of smallest to largest as the heterogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the heterogeneous m-level iterative continuation advertisements of each initial recommended advertisement;

[0093] Arrange homogeneous first-level iterative continuation advertisements or heterogeneous first-level iterative continuation advertisements as target advertisements according to the order of their corresponding target advertisements in the initial recommendation path to form a mixed first-level iterative recommendation path, send the mixed first-level iterative recommendation path to the user terminal, and send the browsing behavior of the user terminal browsing the mixed first-level iterative recommendation path to step 1;

[0094] When the mixed promotion strategy is generated for the second time, the homogeneous secondary iterative continuation advertisements or heterogeneous secondary iterative continuation advertisements as the target advertisements are arranged in the order of their corresponding target advertisements in the initial recommendation path to form a mixed secondary iterative recommendation path, the mixed secondary iterative recommendation path is sent to the user end, and the browsing behavior of the user end browsing the mixed secondary iterative recommendation path is sent to step one; and so on, when the mixed promotion strategy is generated m times, the homogeneous m-level iterative continuation advertisements or heterogeneous m-level iterative continuation advertisements as the target advertisements are arranged in the order of their corresponding target advertisements in the initial recommendation path to form a mixed m-level iterative recommendation path, the mixed m-level iterative recommendation path is sent to the user end, and the browsing behavior of the user end browsing the mixed m-level iterative recommendation path is sent to step one;

[0095] By analyzing the user's interest parameters and browsing behavior to obtain the recommendation value and recommendation tendency value, we can then determine whether to implement homogeneous promotion, heterogeneous promotion or mixed promotion strategies, and iterate to form a dynamically changing promotion strategy, so that the advertisements received by users are more diversified and personalized, improving the user's usage experience and satisfaction; achieving personalized and accurate promotion of electronic components products and improving user satisfaction.

[0096] According to an embodiment of the present invention, there is also provided a method for promoting electronic components products based on big data, the method comprising the following steps:

[0097] R1: Obtain user information and send it to the platform database for storage; the platform database stores all product advertisements on the promotion platform;

[0098] R2: Perform interest analysis based on the user's purchase history and browsing history to obtain the user's interest parameters; perform correlation analysis on the product advertisements on the promotion platform to obtain promotion parameters, and send the interest parameters and promotion parameters to the promotion optimization module; the interest parameters include the relevant product set and the promotion base value corresponding to each tag, and the promotion base value corresponding to the product advertisement included in each tag; the promotion parameters include the correlation relationship between each advertisement and the corresponding relationship index; the correlation relationship is divided into the same type of relationship and the different type of relationship, and the relationship index is divided into the homogeneous index and the correlation index;

[0099] R3: When a user enters the promotion platform through the user terminal, the user's interest parameters are retrieved, and the promotion base value corresponding to each label is multiplied by the set quantity conversion coefficient to obtain the number of advertisements corresponding to each label. The promotion base values ​​corresponding to the product advertisements of each label are selected in descending order until the number of selected product advertisements is consistent with the number of advertisements corresponding to the label; the promotion base value of the label to which the selected product advertisement belongs and the promotion base value corresponding to the product advertisement are retrieved, and the recommended value is obtained by formula calculation and analysis; the initial promotion path is generated for the selected product advertisement in descending order according to the corresponding recommended values, and the selected product advertisement is recorded as the initial recommended product advertisement, and the initial recommended product advertisement is promoted to the user terminal according to the initial promotion path;

[0100] R4: retrieve the user's browsing behavior of the initial recommended product advertisement, and conduct detailed analysis on it to obtain the recommendation tendency value; when the recommendation tendency value is greater than the maximum value in the set tendency interval, execute the homogeneous promotion strategy; when the recommendation tendency is within the set tendency interval, execute the mixed promotion strategy; when the recommendation tendency value is less than the minimum value in the set tendency interval, execute the heterogeneous promotion strategy; after each execution of the corresponding promotion strategy, update the user's browsing behavior to R1 to form a dynamic iteration.

[0101] The above are only embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.

Claims

1. An electronic component product promotion platform based on big data, characterized in that: include: Multi-dimensional analysis module and promotion optimization module; The multidimensional analysis module performs interest analysis based on the user's purchase history and browsing history to obtain the user's interest parameters; at the same time, the product advertisements on the promotion platform are associated with analysis to obtain promotion parameters, and the interest parameters and promotion parameters are sent to the promotion optimization module; the interest parameters include the relevant product set and the promotion base value corresponding to each tag, and the promotion base value corresponding to the product advertisement included in each tag; the promotion parameters include the association relationship between each advertisement and the corresponding relationship index; the association relationship is divided into the same type of relationship and the different type of relationship, and the relationship index is divided into the homogeneous index and the association index; The promotion optimization module analyzes the interest parameters, promotion parameters and user browsing behavior to generate corresponding promotion strategies. The specific process is as follows: When a user enters the promotion platform through the user terminal, the user's interest parameters are retrieved, and the promotion base value corresponding to each label is multiplied by the set quantity conversion coefficient to obtain the number of advertisements corresponding to each label. The promotion base values ​​corresponding to the product advertisements of each label are selected in descending order until the number of selected product advertisements is consistent with the number of advertisements corresponding to the label; the promotion base value of the label to which the selected product advertisement belongs is retrieved and recorded as WL, and the promotion base value LM corresponding to the product advertisement, and the values ​​are converted into the promotion base value LM by the set formula. Calculate and obtain the recommendation value Wd, where d1 and d2 are respectively the set proportional coefficients; generate an initial promotion path for the selected product advertisements in descending order of their corresponding recommendation values, record the selected product advertisements as initial recommended product advertisements, and promote the initial recommended product advertisements to the user end according to the initial promotion path; The browsing behavior of users browsing the initial recommended product advertisements is retrieved and analyzed in detail to obtain the recommendation tendency value; when the recommendation tendency value is greater than the maximum value in the set tendency interval, a homogeneous promotion strategy is implemented; when the recommendation tendency is within the set tendency interval, a mixed promotion strategy is implemented; when the recommendation tendency value is less than the minimum value in the set tendency interval, a heterogeneous promotion strategy is implemented; The specific process of conducting detailed analysis on the browsing behavior of users browsing the initial recommended product advertisements is as follows: Compare and analyze the browsing time with the set time interval to divide the browsing time into primary browsing time, secondary browsing time and tertiary browsing time, and record the initial recommended product advertisements corresponding to the primary browsing time, secondary browsing time and tertiary browsing time as primary initial recommended advertisements, secondary initial recommended advertisements and tertiary initial recommended advertisements respectively; count the number of primary initial recommended advertisements, secondary initial recommended advertisements and tertiary initial recommended advertisements respectively, and record them as B1, B2 and B3 respectively; sum the primary browsing time, secondary browsing time and tertiary browsing time to obtain the primary browsing value, secondary browsing value and tertiary browsing value respectively, and record them as B4, B5 and B6 respectively; The number of first-level initial recommended ads B1, the number of second-level initial recommended ads B2, the number of third-level initial recommended ads B3, the first-level browsing value B4, the second-level browsing value B5 and the third-level browsing value B6 are calculated by the set formula The recommended tendency value Bd is calculated, where d3, d4, and d5 are respectively the set proportional coefficients, and d3>d4>d5>0; The specific process of homogeneous promotion strategy is: Take any initial recommended product advertisement as the target advertisement, retrieve similar advertisements of the target advertisement and their corresponding homogeneity indexes, select the similar advertisement with the largest homogeneity index as the homogeneous first-level iterative continuation advertisement of the target advertisement, select the similar advertisement with the second highest homogeneity index in descending order as the homogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the homogeneous m-level iterative continuation advertisements of each initial recommended advertisement, where m is a positive integer; When a homogeneous promotion strategy is generated for the mth time, the homogeneous m-level iterative continuation advertisements corresponding to each initial recommended product advertisement are arranged in the order of each initial recommended product advertisement in the initial recommendation path to form a homogeneous m-level iterative recommendation path, and sent to the user end; the browsing behavior of the user end browsing the homogeneous m-level iterative recommendation path is re-analyzed for promotion; The specific steps of the heterogeneous promotion strategy are: Take any initial recommended product advertisement as the target advertisement, retrieve the non-identical advertisements of the target advertisement and their corresponding correlation indexes, select the non-identical advertisement with the smallest correlation index as the heterogeneous first-level iterative continuation advertisement of the target advertisement, select the non-identical advertisement with the second highest correlation index in the order of small to large as the heterogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the heterogeneous m-level iterative continuation advertisements of each initial recommended advertisement; When the heterogeneous promotion strategy is generated for the mth time, the heterogeneous m-level iterative continuation advertisements corresponding to the initial recommended product advertisements are arranged in the order of the initial recommended product advertisements in the initial recommendation path to form a heterogeneous m-level iterative recommendation path, and sent to the user end; the browsing lines of the user end browsing the homogeneous m-level iterative recommendation path are re-analyzed for promotion; The specific steps of the mixed promotion strategy are: Randomly select half of the initial recommended advertisements in the initial recommended path as target advertisements for homogeneous promotion, and the other half as target advertisements for heterogeneous promotion; retrieve the same type of advertisements as target advertisements for homogeneous promotion and their corresponding homogeneity indexes, select the same type of advertisement with the largest homogeneity index as the homogeneous first-level iterative continuation advertisement of its target advertisement, select the same type of advertisement with the second highest homogeneity index in descending order as the homogeneous second-level iterative continuation advertisement of its target advertisement, and so on to obtain the homogeneous m-level iterative continuation advertisements of each initial recommended advertisement; When m times of generated mixed promotion strategies are received, the homogeneous m-level iterative continuation advertisements or heterogeneous m-level iterative continuation advertisements that are their target advertisements are arranged in the order of their corresponding target advertisements in the initial recommendation path to form a mixed m-level iterative recommendation path, the mixed m-level iterative recommendation path is sent to the user end, and the browsing behavior of the user end browsing the mixed m-level iterative recommendation path is re-analyzed for promotion.

2. The electronic component product promotion platform based on big data according to claim 1, characterized in that: The specific steps of interest analysis are: Traverse the product advertisements on the platform and classify them according to their product types, and use the product types as labels, thereby obtaining labels for each product promotion; use TF-IDF technology to extract several keywords corresponding to each product advertisement, so that each product advertisement has a classification label and several keywords; Calculate the time difference between the purchase time of each purchase and the current system time to obtain the relevant duration of each purchase, and accumulate the relevant duration to the corresponding purchased product; The time difference between the start time and the end time of each browsing is calculated to obtain the browsing time of each browsing, the time difference between the end time of each browsing and the current time of the system is calculated to obtain the relevant time corresponding to each browsing number, and the browsing time and the relevant time are calculated and analyzed in a formula to obtain the browsing related value corresponding to each browsing number; thereby, each browsing related value is accumulated to the corresponding browsing product; The products purchased and browsed by the user are organized into a set of related products for the user, where each product is recorded as a related product of the user; thus, the cumulative number of purchase related durations of each related product and the corresponding related duration of each accumulation are recorded as Mi, the cumulative number of browsing related values ​​and the corresponding browsing related value of each accumulation are recorded as Lj; Where i=1, 2, 3...I, j=1, 2, 3...J, I and J are all positive integers, I represents the total cumulative number of purchase-related durations, J represents the total cumulative number of browsing-related values, i represents the cumulative number of any purchase-related durations, j represents the cumulative number of any browsing-related values; Mi represents the purchase-related duration corresponding to any cumulative purchase-related duration, Lj represents the browsing-related value corresponding to any cumulative browsing-related value; using the set formula Calculate and obtain the base promotion value LM of each product advertisement, where a3 and a4 are respectively the set proportional coefficients; retrieve the label of each product advertisement, and classify each product advertisement according to the label, so that each label includes several product advertisements, and the sum of the promotion base values ​​of the product advertisements in the same label is calculated as the base promotion value of the label.

3. The electronic component product promotion platform based on big data according to claim 1, characterized in that: The specific steps of association analysis are: Retrieve all product advertisements in the promotion platform, select one of the product advertisements as the target advertisement, and use the other product advertisements as reference advertisements. When the reference advertisement and the target advertisement are in the same label, the reference advertisement is recorded as a similar advertisement; when the reference advertisement and the target advertisement are not in the same label, the reference advertisement is recorded as a different advertisement; count the intersection keywords and union keywords between similar advertisements and target advertisements, and use TF-IDF technology to extract the TF-IDF scores of each intersection keyword in similar advertisements and target advertisements. It can be seen that one intersection keyword corresponds to two TF-IDF scores, namely the TF-IDF score of the intersection keyword in the similar advertisement and the TF-IDF score of the intersection keyword in the target advertisement; sum up the two TF-IDF scores corresponding to the intersection keyword to obtain the intersection score of the intersection keyword; Compare and analyze the intersection score with the set score range to divide the intersection score into a high intersection score, a medium intersection score and a low intersection score, count the number of high intersection scores, medium intersection scores and low intersection scores respectively, and count the number of intersection keywords and union keywords; perform formulaic calculation and analysis on the number of high intersection scores, the number of medium intersection scores, the number of low intersection scores, the number of intersection keywords, the number of union keywords and the intersection score to obtain a homogeneity index between the same type of advertisement and the target advertisement; Retrieve the target advertisement and the product represented by the non-same advertisement, and determine the relationship between the two products, where the relationship is divided into: direct connection relationship, indirect connection relationship, and the order of indirect connection is recorded as n; set a connection coefficient corresponding to each different connection relationship and record it as A1; The number of intersection keywords and union keywords between the target advertisement and the non-same advertisement is counted, and the connection coefficient and the order of indirect connection are analyzed by formula calculation to obtain the correlation index between the non-same advertisement and the target advertisement.

4. A method for promoting electronic components based on big data, characterized in that Applied to an electronic component product promotion platform based on big data as described in any one of claims 1 to 3, the method comprises the following steps: 801: Obtain user information and send it to the platform database for storage; the platform database stores all product advertisements on the promotion platform; 802: performing interest analysis based on the user's purchase history and browsing history to obtain interest parameters of the user; performing correlation analysis on the product advertisements of the promotion platform to obtain promotion parameters, and sending the interest parameters and promotion parameters to the promotion optimization module; wherein the interest parameters include the relevant product set and the promotion base value corresponding to each tag therein, and the promotion base value corresponding to the product advertisement included in each tag; the promotion parameters include the correlation relationship between each advertisement and the corresponding relationship index; the correlation relationship is divided into the same type relationship and the different type relationship, and the relationship index is divided into the homogeneous index and the correlation index; 803: When the user enters the promotion platform through the user terminal, the user's interest parameter is retrieved, the promotion base value corresponding to each tag is multiplied by the set quantity conversion coefficient to obtain the number of advertisements corresponding to each tag, and the promotion base values ​​corresponding to the product advertisements of each tag are selected in descending order until the number of selected product advertisements is consistent with the number of advertisements corresponding to the tag; the promotion base value of the tag to which the selected product advertisement belongs and the promotion base value corresponding to the product advertisement are retrieved, and a formula calculation and analysis are performed to obtain a recommended value; the selected product advertisement is generated into an initial promotion path in descending order according to its corresponding recommended value, and the selected product advertisement is recorded as an initial recommended product advertisement, and the initial recommended product advertisement is promoted to the user terminal according to the initial promotion path; 804: retrieve the browsing behavior of the user browsing the initial recommended product advertisement, and perform detailed analysis on it to obtain the recommendation tendency value; when the recommendation tendency value is greater than the maximum value in the set tendency interval, execute the homogeneous promotion strategy of 805; when the recommendation tendency is within the set tendency interval, execute the mixed promotion strategy of 807; when the recommendation tendency value is less than the minimum value in the set tendency interval, execute the heterogeneous promotion strategy of 806; 805: Homogeneous promotion strategy: Take any initial recommended product advertisement as the target advertisement, retrieve similar advertisements of the target advertisement and their corresponding homogeneity indexes, select the similar advertisement with the largest homogeneity index as the homogeneous first-level iterative continuation advertisement of the target advertisement, select the similar advertisement with the second highest homogeneity index in descending order as the homogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the homogeneous m-level iterative continuation advertisements of each initial recommended advertisement, where m is a positive integer; When a homogeneous promotion strategy is generated for the mth time, the homogeneous m-level iterative continuation advertisements corresponding to the initial recommended product advertisements are arranged in the order of the initial recommended product advertisements in the initial recommendation path to form a homogeneous m-level iterative recommendation path, and the homogeneous m-level iterative recommendation path is sent to the user terminal; the browsing behavior of the user terminal browsing the homogeneous m-level iterative recommendation path is updated to 801; 806: Heterogeneous promotion strategy: Take any initial recommended product advertisement as the target advertisement, retrieve the non-identical advertisements of the target advertisement and their corresponding correlation indexes, select the non-identical advertisement with the smallest correlation index as the heterogeneous first-level iterative continuation advertisement of the target advertisement, select the non-identical advertisement with the second highest correlation index in the order of small to large as the heterogeneous second-level iterative continuation advertisement of the target advertisement, and so on to obtain the heterogeneous m-level iterative continuation advertisements of each initial recommended advertisement; When the heterogeneous promotion strategy is generated for the mth time, the heterogeneous m-level iterative continuation advertisements corresponding to the initial recommended product advertisements are arranged according to the order of the initial recommended product advertisements in the initial recommendation path to form a heterogeneous m-level iterative recommendation path, and sent to the user end; the browsing line of the user end browsing the homogeneous m-level iterative recommendation path is updated to 801; 807: Mixed promotion strategy: Randomly select half of the initial recommended advertisements in the initial recommended path as target advertisements for homogeneous promotion, and the other half as target advertisements for heterogeneous promotion; retrieve the same type of advertisements as target advertisements for homogeneous promotion and their corresponding homogeneity indexes, select the same type of advertisement with the largest homogeneity index as the homogeneous first-level iterative continuation advertisement of its target advertisement, select the same type of advertisement with the second highest homogeneity index in descending order as the homogeneous second-level iterative continuation advertisement of its target advertisement, and so on to obtain the homogeneous m-level iterative continuation advertisements of each initial recommended advertisement; When m times of generating a hybrid promotion strategy are received, the homogeneous m-level iterative continuation advertisements or heterogeneous m-level iterative continuation advertisements as their target advertisements are arranged in the order of their corresponding target advertisements in the initial recommendation path to form a hybrid m-level iterative recommendation path, the hybrid m-level iterative recommendation path is sent to the user end, and the browsing behavior of the user end browsing the hybrid m-level iterative recommendation path is updated to 801.

Images