Information stream advertisement matching system and method based on user browsing and consumption behavior

Abstract

The application discloses an information flow advertisement matching system and method based on user browsing and consumption behaviors, and relates to the technical field of advertisement matching.The matching method comprises the following steps: obtaining any user behavior record, and performing effectiveness evaluation on the user behavior record; based on the effectiveness evaluation result of any user behavior record, a corresponding feature extraction strategy is adopted to perform feature extraction; the influence degree of each extracted feature is analyzed, and a behavior feature evaluation model is established to perform feature evaluation on the user behavior record; the behavior feature evaluation model is adjusted according to the difference presented by the two evaluation results, and a feature threshold value for judging the effectiveness of each user behavior record is obtained; feature information of advertisement information presented by any advertisement is extracted, and the matching effectiveness of the advertisement is evaluated; whether to recommend is judged based on the evaluation result, and each piece of advertisement to be recommended is prioritized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of advertisement matching technology, in particular to an information flow advertisement matching system and method based on user browsing and consumption behavior. BACKGROUND

[0002] Information flow advertisement is a form of advertisement embedded in the user's daily browsing content, usually appearing in social media, news websites or video platforms, etc.; its main feature is highly integrated with the surrounding content, giving users a natural browsing experience;

[0003] On various platforms, the browsing and consumption behavior of users can be collected to filter similar goods for advertisement recommendation to users; under this recommendation mechanism, although the browsing selectivity of users can be increased to a certain extent, there are still problems, because various advertisements will be filled on the platform, and users will easily click on unwanted advertisements when browsing advertisements on various platforms, resulting in invalid browsing records, but the platform cannot distinguish such invalid records, and will continue to push related advertisements based on invalid records, which will affect the browsing and consumption burden of users. SUMMARY

[0004] The purpose of the present application is to provide an information flow advertisement matching system and method based on user browsing and consumption behavior to solve the problems in the prior art.

[0005] To achieve the above purpose, the present application provides the following technical scheme: an information flow advertisement matching method based on user browsing and consumption behavior, the matching method comprising the following steps:

[0006] Step S100: After authorization by the user, the behavior record management system collects the browsing behavior and consumption behavior of the user on all platforms, generates a plurality of user behavior records, and obtains the user feedback recorded in any user behavior record to evaluate the effectiveness of the user behavior record;

[0007] Step S200: Different feature extraction strategies are set for different effectiveness evaluation results, corresponding feature extraction strategies are adopted for feature extraction based on the effectiveness evaluation result of any user behavior record, and the influence degree of each feature is analyzed based on the distribution of each feature between each user behavior record;

[0008] Step S300: Based on each feature extracted from any user behavior record, a behavior feature evaluation model is established to evaluate the features of the user behavior record; the difference between the two evaluation results of the user behavior record is analyzed, the behavior feature evaluation model is adjusted, and a feature threshold for judging the effectiveness of each user behavior record is obtained;

[0009] Step S400: extracting several features in each user behavior record, extracting feature information of the presented advertisement information of any advertisement, and comparing the similarity of the several features in each user behavior record, evaluating the matching effectiveness of the advertisement; judging whether to recommend based on the evaluation result of the advertisement, and prioritizing each advertisement to be recommended.

[0010] Further, step S100 includes the following steps:

[0011] Step S101: whenever a user enters an information detail page through an advertisement presented on any platform, or a user has a consumption behavior, a corresponding user behavior record is generated in the behavior record management system; wherein the user behavior record containing browsing behavior is set as a browsing behavior record, and the user behavior record containing consumption behavior is set as a consumption behavior record;

[0012] Step S102: any browsing behavior record is selected, the information detail page where the user is in the browsing behavior record is obtained, and the browsing data of the user in the information detail page is extracted; a plurality of preset evaluation rules are obtained, the corresponding browsing data is extracted from the browsing data according to each dimension, and the effective value presented under any dimension is obtained, and the effective value of the browsing behavior record is obtained by accumulation; the browsing data includes click rate, dwell time, scroll depth and other data;

[0013] Step S103: any consumption behavior record is selected, the recorded product information in the consumption behavior record is obtained, and the recorded product information in any browsing behavior record is compared; if there are two sets of identical product information, the compared browsing behavior record is marked with a feature; the user's consumption behavior directly indicates that the user's browsing of the corresponding product information is an effective operation, so the effective value of the browsing record of the product corresponding to the consumption behavior can be used as a basis for judging whether each user behavior record is effective, which is beneficial to subsequent effective record differentiation;

[0014] Step S104: obtaining the effective value of each browsing behavior record with a feature mark, selecting the smallest effective value as an effective threshold for judging whether any user behavior record is effective; any user behavior record is selected, if the effective value of the user behavior record exceeds the effective threshold, the user behavior record is set as an effective behavior record, and if it is lower than the effective threshold, it is set as an invalid behavior record.

[0015] Further, step S200 includes the following steps:

[0016] Step S201: Arbitrarily select a user behavior record, if the user behavior record is a valid behavior record, extract the recorded commodity information from the user behavior record, extract several commodity features, and obtain each valid feature of the user behavior record; summarize each valid feature in all valid behavior records to obtain a valid feature set;

[0017] Step S202: If the user behavior record is an invalid behavior record, commodity feature extraction is also performed on the recorded commodity information to obtain several invalid features of the user behavior record; arbitrarily select an invalid feature and any valid feature in the valid feature set for similarity comparison, if the similarity exceeds the set similarity threshold, remove the invalid feature; after invalid feature investigation of each invalid behavior record, generate an invalid feature set; the recorded commodity information in the invalid behavior record does not represent that each feature of the commodity is invalid, because the valid record needs each feature to meet the user's demand, but the invalid record only needs an invalid feature to achieve, so further distinction of each feature in the invalid record is needed to obtain an accurate invalid feature set;

[0018] Step S203: Summarize the valid feature set and the invalid feature set to obtain a feature set, arbitrarily select the i-th feature from the feature set, and obtain the valid feature value of the i-th feature as T i =Ju1(Flag i =1), wherein Ju1() is a judgment function, Flag i is the validity mark of the i-th feature, if the i-th feature is a valid feature, Flag i =1, if the i-th feature is an invalid feature, Flag i =0, if Flag i =1, T i =Ju1(Flag i =1)=1, if Flag i =0, T i =Ju1(Flag i =1)=0;

[0019] Step S204: Obtain the number of valid behavior records containing the i-th feature as m i1 and the number of invalid behavior records containing the i-th feature as m i2 , set the number of valid behavior records stored in the behavior record management system as a and the number of invalid behavior records as b; according to the formula:

[0020] ;

[0021] The influence degree Y of the i-th feature is calculated i The influence degree of the effective feature is a positive promotion to the effective value of each user behavior record, and the influence degree of the ineffective feature is a negative promotion to the effective value, and the appearance frequency of each feature can effectively reflect the influence degree of each feature, which is conducive to subsequent judgment of the effectiveness of each behavior record through the influence of the feature.

[0022] Further, step S300 includes the following steps:

[0023] Step S301: randomly selecting a user behavior record, extracting all features in the user behavior record, wherein the influence degree of the i-th feature is Y i ; and establishing a behavior feature evaluation model:

[0024] ;

[0025] wherein x is the number of features of the user behavior record, Ju() is a judgment function, Record is the effective label of the user behavior record, if the user behavior record is an effective behavior record, then Ju(Record)=1, if it is an ineffective behavior record, then Ju(Record)=-1, k and c are both constant coefficients; and the feature value Z of the user behavior record is calculated;

[0026] Step S302: extracting the feature value Z' of any one effective behavior record and the feature value Z'' of any one ineffective behavior record, respectively, if Z'<Z'', then adjusting the constant coefficients k and c so that Z'≥Z'', training the behavior feature evaluation model so that the feature value of any effective behavior record is greater than the feature value of any ineffective behavior record, thereby determining the numerical value of the constant coefficients k and c;

[0027] Step S303: obtaining the maximum feature value Z ’’ max in all ineffective behavior records and the minimum feature value Z ’ min in all effective behavior records, obtaining the feature threshold Z th =(Z ’’ max +Z ’ min ) / 2 for judging whether each user behavior record is effective.

[0028] In order to better realize the above method, an information flow advertisement matching system is also proposed, which includes a user behavior screening module, a feature information processing module, a user behavior evaluation module and a real-time advertisement pushing module.

[0029] The user behavior screening module is configured to collect browsing behavior and consumption behavior of a user on all platforms after the user is authorized, generate a plurality of user behavior records, and obtain user feedback recorded in any user behavior record to evaluate the effectiveness of the user behavior record.

[0030] The feature information processing module is configured to set different feature extraction strategies for different effectiveness evaluation results, perform feature extraction based on the effectiveness evaluation result of any user behavior record, and analyze the influence degree of each feature based on the distribution of each feature between the user behavior records.

[0031] The user behavior evaluation module is configured to establish a behavior feature evaluation model based on each feature extracted from any user behavior record, perform feature evaluation on the user behavior record, analyze the difference between the two evaluation results of the user behavior record, adjust the behavior feature evaluation model, and obtain a feature threshold for judging the effectiveness of each user behavior record.

[0032] The real-time advertisement pushing module is configured to extract a plurality of features from each user behavior record, extract feature information from advertisement information presented by any advertisement, compare the similarity between the feature information and the plurality of features in each user behavior record, evaluate the matching effectiveness of the advertisement, determine whether to recommend the advertisement based on the evaluation result of the advertisement, and prioritize each advertisement to be recommended.

[0033] Further, the user behavior screening module includes a behavior record setting unit and an effective behavior evaluation unit.

[0034] The behavior record setting unit is configured to collect browsing behavior and consumption behavior of a user on all platforms after the user is authorized, generate a plurality of user behavior records, and the effective behavior evaluation unit is configured to obtain user feedback recorded in any user behavior record to evaluate the effectiveness of the user behavior record.

[0035] Further, the feature information processing module includes a feature information extraction unit and an influence degree calculation unit.

[0036] The feature information extraction unit is configured to set different feature extraction strategies for different effectiveness evaluation results, perform feature extraction based on the effectiveness evaluation result of any user behavior record, and the influence degree calculation unit is configured to analyze the influence degree of each feature based on the distribution of each feature between the user behavior records.

[0037] Further, the user behavior evaluation module comprises a behavior feature evaluation unit and a feature threshold value determination unit;

[0038] The feature evaluation unit is configured to establish a behavior feature evaluation model based on the features extracted from the user behavior records, and to evaluate the user behavior records based on the behavior feature evaluation model.

[0039] Further, the real-time advertisement pushing module comprises an advertisement matching evaluation unit and a pushing sequence setting unit.

[0040] The advertisement matching evaluation unit is configured to extract features from the user behavior records, to extract feature information from the advertisement information of any advertisement, to compare the similarity between the feature information and the features of the user behavior records, and to evaluate the matching effectiveness of the advertisement.

[0041] Compared with the prior art, the present application has the following advantages:

[0042] 1. The present application can distinguish the behavior records of the user, help to exclude invalid records, provide more accurate advertisement recommendation for the user, improve the browsing selectivity of the user, and effectively reduce the browsing burden of the user.

[0043] 2. The present application can deeply mine the expected product features of the user by extracting features from the product information and performing effectiveness analysis, can more accurately recommend the advertisements, can provide more accurate recommendation suggestions on the basis of the original conventional recommendation strategy, and can check the advertisements that do not meet the actual requirements, thereby reducing the browsing burden of the user. BRIEF DESCRIPTION OF DRAWINGS

[0044] Fig. 1 FIG. 1 is a step schematic diagram of the information flow advertisement matching method based on the browsing and consumption behavior of the user.

[0045] Fig. 2 FIG. 2 is a structure schematic diagram of the information flow advertisement matching system based on the browsing and consumption behavior of the user. DETAILED DESCRIPTION

[0046] Clearly, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0047] Embodiment: As shown in the figure, the present application provides an information flow advertisement matching method based on user browsing and consumption behavior, and the matching method comprises the following steps: Figs. 1-2

[0048] Step S100: After the user authorization, the behavior record management system collects the user's browsing behavior and consumption behavior on all platforms, generates a plurality of user behavior records, and obtains the user feedback recorded in any user behavior record to evaluate the effectiveness of the user behavior record.

[0049] Among them, step S100 comprises the following steps:

[0050] Step S101: Whenever the user enters the information detail page through the advertisement presented on the platform on any platform, or the user has a consumption behavior, a corresponding user behavior record is generated in the behavior record management system; wherein the user behavior record containing browsing behavior is set as browsing behavior record, and the user behavior record containing consumption behavior is set as consumption behavior record;

[0051] Step S102: Any browsing behavior record is selected, the information detail page where the user is in the browsing behavior record is obtained, the browsing data of the user in the information detail page is extracted, a plurality of dimension evaluation rules are preset, the corresponding browsing data is extracted from the browsing data according to each dimension, and the effective value presented under any dimension is obtained, and the effective value of the browsing behavior record is obtained by accumulation;

[0052] Step S103: Any consumption behavior record is selected, the product information recorded in the consumption behavior record is obtained, the product information is compared with the product information recorded in any browsing behavior record, if there are two groups of same product information, the compared browsing behavior record is marked with a feature;

[0053] Step S104: The effective value of each browsing behavior record with feature mark is obtained, the smallest effective value is selected as the effective threshold value for judging whether any user behavior record is effective; any user behavior record is selected, if the effective value of the user behavior record exceeds the effective threshold value, the user behavior record is set as effective behavior record, if it is lower than the effective threshold value, it is set as invalid behavior record. ​

[0054] Step S200: setting different feature extraction strategies for different effectiveness evaluation results, based on the effectiveness evaluation results of any user behavior record, taking the corresponding feature extraction strategy to perform feature extraction; based on the distribution of each feature between each user behavior record, analyzing the influence degree of each extracted feature;

[0055] Among them, step S200 includes the following steps:

[0056] Step S201: randomly selecting a user behavior record, if the user behavior record is an effective behavior record, extracting the recorded product information from the user behavior record, extracting several product features, and obtaining each effective feature of the user behavior record; all effective features in all effective behavior records are summarized to obtain an effective feature set;

[0057] Step S202: if the user behavior record is an invalid behavior record, the recorded product information is also subjected to product feature extraction to obtain several invalid features of the user behavior record; any invalid feature is compared with any effective feature in the effective feature set for similarity, if the similarity exceeds the set similarity threshold, the invalid feature is removed; after invalid feature investigation of each invalid behavior record, an invalid feature set is generated;

[0058] Step S203: the effective feature set and the invalid feature set are summarized to obtain a feature set, the i-th feature is randomly selected from the feature set, and the effective feature value of the i-th feature is T i =Ju1(Flag i =1), wherein Ju1() is a judgment function, Flag i is the effectiveness mark of the i-th feature, if the i-th feature is an effective feature, Flag i =1, if the i-th feature is an invalid feature, Flag i =0, if Flag i =1, then T i =Ju1(Flag i =1)=1, if Flag i =0, then T i =Ju1(Flag i =1)=0;

[0059] Step S204: obtaining the number of effective behavior records containing the i-th feature m i1 and the number of invalid behavior records containing the i-th feature m i2 , setting the number of effective behavior records stored in the behavior record management system as a and the number of invalid behavior records as b; according to the formula:

[0060] ;

[0061] The influence degree Y of the ith feature is calculated i ;

[0062] In the embodiment 1, the number of valid behavior records containing the ith feature is set as 10, the number of invalid behavior records is set as 5, the number of valid behavior records stored in the behavior record management system is set as 40, and the number of invalid behavior records is set as 10. The ith feature is set as a valid feature, and T is obtained as 1, Y is obtained as 10 / 15x(1+1 / 4-1 / 2)x1=50%. i i

[0063] Step S300: based on each feature extracted from the arbitrary user behavior record, a behavior feature evaluation model is established to evaluate the user behavior record, the difference between the two evaluation results of the user behavior record is analyzed, the behavior feature evaluation model is adjusted, and the feature threshold for judging the validity of each user behavior record is obtained.

[0064] The step S300 includes the following steps:

[0065] Step S301: an arbitrary user behavior record is selected, and all features in the user behavior record are extracted, wherein the influence degree of the ith feature is set as Y i , and a behavior feature evaluation model is established:

[0066] ;

[0067] wherein x is the number of features of the user behavior record, Ju() is a judgment function, Record is the valid mark of the user behavior record, if the user behavior record is a valid behavior record, Ju(Record)=1, if the user behavior record is an invalid behavior record, Ju(Record)=-1, k and c are constant coefficients; the feature value Z of the user behavior record is calculated.

[0068] In the embodiment 2, there are three features in the user behavior record, and the influence degree of each feature is 70%, 50% and-20% respectively, k=50%, c=0, and the feature value of the user behavior record is Z=1.7x1.5x0.8-0.5=1.54.

[0069] ​​Step S302: Extract the feature value Z' of any one valid behavior record and the feature value Z'' of any one invalid behavior record respectively, if Z' < Z'', adjust the constant coefficients k and c so that Z' >= Z'', train the behavior feature evaluation model so that the feature value of any valid behavior record is greater than the feature value of any invalid behavior record, thereby determining the numerical value of the constant coefficients k and c;

[0070] Step S303: Obtain the maximum feature value Zmax in all invalid behavior records ’’ max And the minimum feature value Zmin in all valid behavior records ’ min Obtain the feature threshold Zth of judging whether each user behavior record is valid th =(Z ’’ max +Z ’ min ) / 2.

[0071] Step S400: Extract several features in each user behavior record, extract the feature information of the advertisement information presented by any advertisement, and compare the similarity with the several features in each user behavior record, evaluate the matching validity of the advertisement; judge whether to recommend based on the evaluation result of the advertisement, and prioritize each advertisement to be recommended.

[0072] The information flow advertisement matching system includes a user behavior screening module, a feature information processing module, a user behavior evaluation module, and a real-time advertisement pushing module.

[0073] The user behavior screening module is used to build a behavior record management system, which collects the browsing behavior and consumption behavior of users on all platforms after user authorization, generates several user behavior records, and obtains the user feedback recorded in any user behavior record to evaluate the validity of the user behavior record.

[0074] The feature information processing module is used to set different feature extraction strategies for different validity evaluation results, extract features based on the validity evaluation results of any user behavior record, and analyze the influence degree of each feature based on the distribution of each feature between each user behavior record.

[0075] The user behavior evaluation module is configured to establish a behavior feature evaluation model based on the features extracted from the user behavior records, to evaluate the user behavior records, to analyze the differences between the evaluation results of the user behavior records, to adjust the behavior feature evaluation model, and to obtain the feature threshold for judging the effectiveness of the user behavior records.

[0076] The real-time advertisement pushing module is configured to extract features from the user behavior records, to extract feature information from the advertisement information of any advertisement, to compare the features of the user behavior records with the feature information of the advertisement, to evaluate the matching effectiveness of the advertisement, and to determine whether to recommend the advertisement based on the evaluation result of the advertisement and to prioritize the recommended advertisements.

[0077] The user behavior screening module includes a behavior record setting unit and an effective behavior evaluation unit.

[0078] The behavior record setting unit is configured to collect the browsing behavior and consumption behavior of a user on all platforms after the user authorizes the behavior record management system, to generate user behavior records, and the effective behavior evaluation unit is configured to obtain the user feedback recorded in any user behavior record, to evaluate the effectiveness of the user behavior record.

[0079] The feature information processing module includes a feature information extraction unit and an influence degree calculation unit.

[0080] The feature information extraction unit is configured to set different feature extraction strategies for different effectiveness evaluation results, to extract features based on the effectiveness evaluation results of the user behavior records, and the influence degree calculation unit is configured to analyze the influence degree of the extracted features based on the distribution of the features among the user behavior records.

[0081] The user behavior evaluation module includes a behavior feature evaluation unit and a feature threshold determination unit.

[0082] The feature evaluation unit is configured to establish a behavior feature evaluation model based on the features extracted from the user behavior records, to evaluate the user behavior records, and the feature threshold determination unit is configured to analyze the differences between the evaluation results of the user behavior records, to adjust the behavior feature evaluation model, and to obtain the feature threshold for judging the effectiveness of the user behavior records.

[0083] The real-time advertisement pushing module includes an advertisement matching evaluation unit and a pushing sequence setting unit.

[0084] An advertisement matching evaluation unit is configured to extract a plurality of features in each user behavior record, extract feature information from advertisement information presented by any advertisement, compare the feature information with the plurality of features in each user behavior record, and evaluate matching effectiveness of the advertisement; and a push sequence setting unit is configured to determine whether to recommend the advertisement based on the evaluation result of the advertisement, and prioritize each advertisement to be recommended.

[0085] It will be apparent to those skilled in the art that the application is not limited to the details of the above-exemplified embodiments and that the application can be implemented in other particular forms without departing from the spirit or essential characteristics of the application. The embodiments should therefore be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the above description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. No reference signs in the claims should be considered as limiting the scope of the claims with respect to the figures of the application.

Claims

1. A method for matching information flow advertisements based on user browsing and consumption behavior, characterized in that: The matching method includes the following steps: Step S100: After user authorization, the behavior record management system collects the user's browsing and consumption behavior on all platforms and generates several user behavior records; obtains user feedback recorded in any user behavior record and evaluates the effectiveness of the user behavior records. Step S200: Set different feature extraction strategies for different validity evaluation results. Based on the validity evaluation results of any user behavior record, adopt the corresponding feature extraction strategy to extract features. Based on the distribution of each feature among each user behavior record, analyze the degree of influence of each extracted feature. Step S300: Based on the features extracted from any user behavior record, establish a behavior feature evaluation model to evaluate the user behavior record; analyze the differences between the two evaluation results of the user behavior record, adjust the behavior feature evaluation model, and obtain the feature threshold for judging the validity of each user behavior record. Step S400: Extract several features from each user behavior record, extract feature information from the advertising information presented by any advertisement, and compare the similarity with several features in each user behavior record to evaluate the matching effectiveness of the advertisement; determine whether to recommend the advertisement based on the evaluation result of the advertisement, and prioritize the recommended advertisements. Step S300 includes the following steps: Step S301: Randomly select a user behavior record and extract all features from the user behavior record, wherein the influence degree of the i-th feature is set as Y. i Establish a behavioral characteristic assessment model: ； Where x is the number of features of the user behavior record, Ju() is the judgment function, Record is the valid marker of the user behavior record, if the user behavior record is a valid behavior record, then Ju(Record)=1, if it is an invalid behavior record, then Ju(Record)=-1, k and c are constant coefficients; the feature value Z of the user behavior record is calculated. Step S302: Extract the feature value Z' of any valid behavior record and the feature value Z'' of any invalid behavior record respectively. If Z' < Z'', adjust the constant coefficients k and c so that Z' ≥ Z''. Train the behavior feature evaluation model so that the feature value of any valid behavior record is greater than the feature value of any invalid behavior record, thereby determining the values ​​of the constant coefficients k and c. Step S303: Obtain the feature value with the largest value among all invalid behavior records, Z. ’’ max The feature with the smallest value among all valid behavior records is Z. ’ min The feature threshold Z for determining whether each user behavior record is valid is obtained. th =(Z ’’ max +Z ’ min ) / 2.

2. The information flow advertising matching method based on user browsing and consumption behavior according to claim 1, characterized in that: Step S100 includes the following steps: Step S101: Whenever a user enters the information details page through an advertisement presented on any platform, or when a user makes a purchase, a corresponding user behavior record is generated in the behavior record management system; user behavior records containing browsing behavior are defined as browsing behavior records, and user behavior records containing purchase behavior are defined as purchase behavior records. Step S102: Randomly select a browsing behavior record, obtain the information details page where the user is located in the browsing behavior record, extract the browsing data of the user in the information details page; preset several evaluation rules for dimensions, extract the corresponding browsing data from the browsing data according to each dimension, and obtain the effective value presented under any dimension, and accumulate them to obtain the effective value of the browsing behavior record; Step S103: Randomly select a consumption behavior record, obtain the product information recorded in the consumption behavior record, compare the product information with the product information recorded in any browsing behavior record, and if there are two sets of the same product information, mark the browsing behavior record being compared with features. Step S104: Obtain the valid values ​​of each browsing behavior record with feature markers, and select the smallest valid value as the valid threshold for judging whether any user behavior record is valid; arbitrarily select a user behavior record, and if the valid value of the user behavior record exceeds the valid threshold, then set the user behavior record as a valid behavior record; if it is lower than the valid threshold, then set it as an invalid behavior record.

3. The information flow advertising matching method based on user browsing and consumption behavior according to claim 2, characterized in that: Step S200 includes the following steps: Step S201: Randomly select a user behavior record. If the user behavior record is a valid behavior record, extract the recorded product information from the user behavior record, extract several product features, and obtain each valid feature of the user behavior record; summarize each valid feature in all valid behavior records to obtain a set of valid features. Step S202: If the user behavior record is an invalid behavior record, the product features of the recorded product information are also extracted to obtain several invalid features of the user behavior record; an invalid feature is randomly selected and compared with any valid feature in the set of valid features. If the similarity exceeds the set similarity threshold, the invalid feature is removed; after checking the invalid features of each invalid behavior record, an invalid feature set is generated. Step S203: Summarize the effective feature set and the invalid feature set to obtain a feature set. Randomly select the i-th feature from the feature set to obtain the effective feature value T of the i-th feature. i =Ju1(Flag i =1), where Ju1() is the judgment function, Flag i Flag is used to mark the validity of the i-th feature. If the i-th feature is a valid feature, then Flag... i =1, if the i-th feature is an invalid feature, then Flag i =0, if Flag i =1, then T i =Ju1(Flag i =1)=1, if Flag i =0, then T i =Ju1(Flag i =1)=0; Step S204: Obtain the number of valid behavior records containing the i-th feature, which is m. i1 The number of invalid behavior records containing the i-th feature is m. i2 Let 'a' be the number of valid behavior records and 'b' be the number of invalid behavior records stored in the behavior record management system; according to the formula: ； The influence degree Y of the i-th feature is calculated. i .

4. An information flow advertising matching system, used to execute the information flow advertising matching method based on user browsing and consumption behavior as described in any one of claims 1-3, characterized in that: The matching system includes a user behavior filtering module, a feature information processing module, a user behavior evaluation module, and a real-time advertising push module. The user behavior filtering module is used to build a behavior record management system that, after user authorization, collects user browsing and consumption behavior on all platforms, generates several user behavior records, obtains user feedback recorded in any user behavior record, and evaluates the effectiveness of the user behavior records. The feature information processing module is used to set different feature extraction strategies for different validity evaluation results. Based on the validity evaluation results of any user behavior record, it adopts the corresponding feature extraction strategy to extract features. Based on the distribution of each feature among the user behavior records, it analyzes the degree of influence of each extracted feature. The user behavior evaluation module is used to establish a behavior feature evaluation model based on the features extracted from any user behavior record to evaluate the user behavior record; analyze the differences between the two evaluation results of the user behavior record, adjust the behavior feature evaluation model, and obtain the feature threshold for judging the validity of each user behavior record. The real-time ad push module is used to extract several features from each user behavior record, extract feature information from the ad information presented by any ad, compare the similarity with several features in each user behavior record, evaluate the matching effectiveness of the ad, determine whether to recommend the ad based on the evaluation result, and prioritize the recommended ads.

5. The information flow advertising matching system according to claim 4, characterized in that: The user behavior filtering module includes a behavior recording setting unit and an effective behavior evaluation unit; The behavior recording setting unit is used to construct a behavior recording management system that, after user authorization, collects user browsing and consumption behaviors on all platforms and generates several user behavior records; the effective behavior evaluation unit is used to obtain user feedback recorded in any user behavior record and evaluate the effectiveness of the user behavior record.

6. The information flow advertising matching system according to claim 4, characterized in that: The feature information processing module includes a feature information extraction unit and an influence degree calculation unit; The feature information extraction unit is used to set different feature extraction strategies for different validity evaluation results, and to extract features based on the validity evaluation results of any user behavior record by adopting the corresponding feature extraction strategy; the influence degree calculation unit is used to analyze the influence degree of each extracted feature based on the distribution of each feature among each user behavior record.

7. The information flow advertising matching system according to claim 4, characterized in that: The user behavior evaluation module includes a behavior feature evaluation unit and a feature threshold determination unit; The feature evaluation unit is used to establish a behavior feature evaluation model based on the features extracted from any user behavior record to evaluate the user behavior record; the feature threshold determination unit is used to analyze the differences between the two evaluation results of the user behavior record, adjust the behavior feature evaluation model, and obtain the feature threshold for judging the validity of each user behavior record.

8. The information flow advertising matching system according to claim 4, characterized in that: The real-time advertising push module includes an advertising matching evaluation unit and a push order setting unit; The ad matching evaluation unit is used to extract several features from each user behavior record, extract feature information from the ad information presented by any ad, and compare the similarity with several features in each user behavior record to evaluate the matching effectiveness of the ad; the push order setting unit is used to determine whether to recommend the ad based on the evaluation result of the ad, and prioritize the recommended ads.

Images