Consumer behavior prediction platform based on big data

Abstract

The invention discloses a consumer behavior prediction platform based on big data, which comprises a data management module, a behavior prediction module and a warning module, and is characterized in that the data management module is used for collecting consumer behavior data and commodity information data, preprocessing the data, storing the data into a database, and extracting parallelism features; and the behavior prediction module is used for analyzing the purchase desire of the consumer by using the historical consumption behavior data of the consumer, screening out months with high purchase desire, analyzing the emotion of the consumer by calling the behavior data of the consumer to obtain a consumer behavior prediction result, and determining the behavior of the consumer according to the prediction result. The consumer behavior data is used for acquiring consumer behavior data, the matching algorithm is used for screening out commodities possibly interested by the consumer and pushing the commodities to the consumer, and the warning system is used for analyzing the deviation degree of the consumer behavior data and historical consumer behavior data, judging whether the consumer behavior is abnormal or not and warning the consumer for the abnormality. The method has the characteristic of improving prediction accuracy and safety.

Description

Technical Field

[0001] The present invention relates to the technical field of moisture detection, and in particular to a consumer behavior prediction platform based on big data. Background Art

[0002] A consumer behavior prediction platform is a tool based on big data, machine learning, and artificial intelligence technologies that is used to analyze and predict consumer purchasing behaviors, preferences, and trends. Such a platform can better help businesses better understand their customers, optimize marketing strategies, and enhance user experience, thereby increasing sales and customer loyalty.

[0003] Existing consumer behavior prediction platforms mostly rely on historical transaction data and simple model predictions, and are unable to handle diverse and multimodal big data. Existing technologies generally combine big data analysis and machine learning with consumer behavior prediction platforms to solve the problem. However, existing platforms ignore the impact of consumer emotions on the accuracy of consumer behavior predictions, and consumers may suffer from consumer fraud when making purchases, resulting in economic losses and consumer safety cannot be guaranteed. Therefore, it is necessary to design a consumer behavior prediction platform based on big data to improve prediction accuracy and security. Summary of the invention

[0004] The purpose of the present invention is to provide a consumer behavior prediction platform based on big data to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a consumer behavior prediction platform based on big data, comprising:

[0006] Data management module, used to manage the data collected by the consumer behavior prediction platform;

[0007] Behavior prediction module, used to predict and evaluate consumers' purchasing behavior and recommend satisfactory products to consumers;

[0008] The warning module is used to detect abnormal consumer behavior and warn consumers.

[0009] According to the above technical solution, the data management module includes:

[0010] A data collection unit, used to collect consumer behavior data and product information data;

[0011] The data processing unit is used to preprocess the collected data and extract corresponding features.

[0012] According to the above technical solution, the data management module includes:

[0013] A consumption desire analysis unit is used to analyze consumers' consumption desire through consumer behavior data and make predictions, and to make consumer behavior predictions in months when consumers' consumption desire is medium or high;

[0014] The emotion prediction unit is used to analyze the emotion of consumers by analyzing the consumer behavior data, and predict the consumer behavior by the influence of emotion on the consumer behavior;

[0015] The product recommendation unit is used to analyze consumers' purchasing behavior and recommend related products to consumers.

[0016] According to the above technical solution, the warning module includes:

[0017] Abnormal behavior detection unit, used to analyze consumers’ purchasing behavior and detect whether the behavior is abnormal;

[0018] The warning unit is used to warn consumers when their behavior is abnormal.

[0019] According to the above technical solution, the method for operating a consumer behavior prediction platform based on big data is characterized in that the method steps specifically include:

[0020] Step S1: Establish a database, collect consumer behavior data and product information data in real time, store the data in the database, and import historical consumer behavior data into the database;

[0021] Step S2: calling the consumer behavior data in the database, preprocessing the data, and extracting features. By analyzing the feature data, the consumer's consumption desire parameters are obtained. According to the consumption desire parameters, the consumer's emotions are analyzed to predict the consumer's consumption behavior.

[0022] Step S3: Based on the results of consumer behavior prediction, when the system determines that the consumer is likely to make a purchase, the system automatically pushes relevant product information to the consumer;

[0023] Step S4: While predicting consumer behavior, analyze the possible purchasing behavior of consumers to determine whether the consumer's purchasing behavior is abnormal and issue a warning.

[0024] According to the above technical solution, establishing a database to collect real-time consumer transaction data includes the user's transaction records, social media data, and mobile application behavior data. The user's transaction records include transaction time, transaction location, amount, and commodity type. The social media data includes comments, likes, and share data of commodities. The mobile application behavior data includes click, browse, and purchase data of commodities. After preprocessing the collected data, it is stored in the database, and historical consumer behavior data is imported into the system;

[0025] Establish a commodity database, extract the features of commodity information data including commodity category, brand, and price. Different commodity categories and brands have corresponding codes, and the code does not exceed 1. And normalize the price, and store the feature data in the commodity database.

[0026] According to the above technical solution, step S2 further includes the following steps:

[0027] Step S21: Call the historical consumer behavior data in the database, preprocess the data, and extract the features in the data including consumption frequency F i 、average consumption amount A i , where i = 1, 2,..., 12, calculate the consumption amount growth rate Calculate the consumer's consumption desire parameter D i = ω F F i + ω A A i + ω R R i , where ω F 、ω A 、ω R are the weight data set in the database, and save the consumption desire parameter in the database according to time;

[0028] Step S22: Call the consumption desire parameter data D i in the database, call the set consumption desire parameter thresholds D1, D2 in the database, D1 < D2. When D i ≤ D1, it is judged that the consumer's consumption desire in this month is low. When D2 > D i > D1, it is judged that the consumer's consumption desire in this month is medium. When D2 ≤ D i , it is judged that the consumer's consumption desire in this month is high. Based on the judgment result, predict the consumer's consumption behavior in the following months when the consumer's consumption desire is high and medium.

[0029] According to the above technical solution, step S22 further includes the following steps:

[0030] Step S221: call the social media data in the database including comments, likes data and mobile application behavior data including clicks, browsing and purchase data of merchants, pre-process the data to remove duplicate values, segment the data, and obtain the call data S = {ω1, ω2, ..., ω i}, where i = 1, 2, ..., n, ω i For a single word, for each word ω i Embed a vector v i , get the data vector Building a sentiment dictionary Where x = 1, 2, ..., γ, γ is the number of words in the sentiment dictionary, For a single word, for each word Embed a vector ε x , get the sentiment dictionary vector And perform sentiment scoring on each word to get the corresponding sentiment score τ x ;

[0031] Step S222: By calculating the single word vector v in the data vector i With the sentiment dictionary vector ε S Similarity By comparing, the sentiment score of the vector with the smallest similarity is selected as the sentiment score τ of a single data vector i , the sentiment score of the data vector is Call the thresholds μ1 and μ2 set in the database, μ1>μ2. When the sentiment score satisfies τ≤μ2, the sentiment of the data vector is classified as negative sentiment. When the sentiment score satisfies μ1>τ>μ2, the sentiment of the data vector is classified as neutral sentiment. When the sentiment score satisfies μ1≤τ, the sentiment of the data vector is classified as positive sentiment, and the corresponding sentiment score τ is output;

[0032] Step S223: Convert the sentiment score τ result into the characteristic data of consumer behavior, and obtain the characteristic data combined with the sentiment score as f=[r,τ], where r is the characteristic value of the corresponding consumer behavior data. Input the characteristic data into the logistic regression algorithm, and predict the probability of consumers making purchase decisions through the regression logic algorithm. where w T The weight set in the database is used to call the threshold δ1 set in the database. When v≥δ1, the consumer makes a purchase, and when δ<δ1, the consumer does not make a purchase.

[0033] According to the above technical solution, step S3 further includes the following steps:

[0034] Step S31: call the feature data in the commodity database including commodity category x i 、Brandyi 、Price i , we get the eigenvector N = [x i ,y i ,z i ], when the system determines that the consumer may make a purchase, the system collects data on the goods that the consumer may purchase, pre-processes the data, and extracts features including product category x, brand y, and price z to obtain a feature vector M = [x, y, z];

[0035] Step S32: Calculate the distance between the product feature vector M and the feature vector N of the product in the product database to obtain the similarity Where n is the number of products in the product database, the threshold d0 set in the database is called, and the products in the product database that meet d i Products with value ≥d0 are pushed to consumers.

[0036] According to the above technical solution, step S4 further includes the following steps:

[0037] Step S41: call the user transaction records in the consumer behavior database including transaction time and amount, preprocess the data experience, and extract features including consumption frequency K and consumption amount distribution Q, call the user transaction record features in the historical consumer behavior database including consumption frequency K i 、Consumption amount distribution Q i , normalize the data;

[0038] Step S42: By formula Calculate the degree of deviation α between the user transaction record characteristics and the historical user transaction record characteristics K , α Q , where X is the type of user transaction record feature, is the average value of historical user transaction record data, σ is the standard deviation of historical user transaction record data, and the total deviation degree is obtained by calculation where ω K ,ω Q For the corresponding weight, the threshold α0 set in the database is called. When α≥α0, the system determines that the consumer behavior is abnormal and there is consumer fraud, and the system consumer is warned.

[0039] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention collects consumer behavior data and product information data, pre-processes the data, stores the data in a database, extracts features in parallel, analyzes historical consumer behavior data, calculates consumer consumption desire parameters, screens out months with high consumption desire for consumer behavior prediction, and uses a matching algorithm to screen out products that consumers may be interested in based on their consumption behavior, and pushes the products to consumers; by providing consumer sentiment analysis and performing consumer behavior prediction, the present invention determines whether consumer behavior is abnormal based on the degree of deviation between consumer behavior data and historical consumer behavior data, thereby improving prediction accuracy and security. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0041] Figure 1 A schematic diagram of the module composition of a consumer behavior prediction platform based on big data provided in Example 2 of the present invention.

[0042] Figure 2 A flowchart of the steps of the operation method of the consumer behavior prediction platform based on big data provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] Embodiment 1: Embodiment 1 of the present invention provides a consumer behavior prediction platform based on big data. Figure 1 A schematic diagram of the module composition of a consumer behavior prediction platform based on big data provided in the second embodiment of the present invention, such as Figure 1 As shown, the system includes:

[0045] Data management module, used to manage the data collected by the consumer behavior prediction platform;

[0046] Behavior prediction module, used to predict and evaluate consumers' purchasing behavior and recommend satisfactory products to consumers;

[0047] The warning module is used to detect abnormal consumer behavior and warn consumers.

[0048] In some embodiments of the present invention, comprising:

[0049] A data collection unit, used to collect consumer behavior data and product information data;

[0050] The data processing unit is used to pre-process the collected data and extract corresponding features.

[0051] In some embodiments of the present invention, comprising:

[0052] A consumption desire analysis unit is used to analyze consumers' consumption desire through consumer behavior data and make predictions, and to make consumer behavior predictions in months when consumers' consumption desire is medium or high;

[0053] The emotion prediction unit is used to analyze the emotion of consumers by analyzing the consumer behavior data, and predict the consumer behavior by the influence of emotion on the consumer behavior;

[0054] The product recommendation unit is used to analyze consumers' purchasing behavior and recommend related products to consumers.

[0055] In some embodiments of the present invention, comprising:

[0056] Abnormal behavior detection unit, used to analyze consumers’ purchasing behavior and detect whether the behavior is abnormal;

[0057] The warning unit is used to warn consumers when their behavior is abnormal.

[0058] Embodiment 2: This embodiment can be applied to the scenario of moisture detection. The method can be executed by the consumer behavior prediction platform based on big data provided by Embodiment 1. The method specifically includes the following steps:

[0059] Step S1: Establish a database, collect consumer behavior data and product information data in real time, store the data in the database, and import historical consumer behavior data into the database;

[0060] Step S2: calling the consumer behavior data in the database, preprocessing the data, and extracting features. By analyzing the feature data, the consumer's consumption desire parameters are obtained. According to the consumption desire parameters, the consumer's emotions are analyzed to predict the consumer's consumption behavior.

[0061] Step S3: Based on the results of consumer behavior prediction, when the system determines that the consumer is likely to make a purchase, the system automatically pushes relevant product information to the consumer;

[0062] Step S4: While predicting consumer behavior, analyze the possible purchase behaviors of consumers to determine whether there are any abnormalities in their purchase behaviors and issue warnings.

[0063] In an embodiment of the present invention, when establishing the database, the real-time collection of consumer transaction data includes the user's transaction records, social media data, and mobile application behavior data. The user's transaction records include transaction time, transaction location, amount, and commodity type. The social media data includes comments, likes, and share data of commodities. The mobile application behavior data includes click, browse, and purchase data of commodities. After preprocessing the collected data, it is stored in the database, and historical consumer behavior data is imported into the system.

[0064] Establish a commodity database, and extract the characteristics of commodity information data including commodity category, brand, and price. Different commodity categories and brands have corresponding codes, and the codes do not exceed 1. Normalize the price and store the feature data in the commodity database.

[0065] In an embodiment of the present invention, step S2 further includes the following steps:

[0066] Step S21: Call the historical consumer behavior data in the database, preprocess the data, and extract the features in the data including the consumption frequency F i , average consumption amount A i , where i = 1, 2,..., 12, and calculate the consumption amount growth rate Calculate the consumer's consumption desire parameter D i = ω F F i + ω A A i + ω R R i , where ω F , ω A , ω R are the weight data set in the database, and save the consumption desire parameter by time in the database;

[0067] Step S22: Call the consumption desire parameter data D i in the database, call the set consumption desire parameter thresholds D1, D2 in the database, D1 < D2. When D i ≤ D1, it is judged that the consumer's consumption desire in this month is low. When D2 > D i > D1, it is judged that the consumer's consumption desire in this month is medium. When D2 ≤ D i , it is judged that the consumer's consumption desire in this month is high. Based on the judgment result, predict the consumer's consumption behavior in the following months when the consumer's consumption desire is high and medium.

[0068] In the embodiment of the present invention, step S22 further includes the following steps:

[0069] Step S221: call the social media data in the database including comments, likes data and mobile application behavior data including clicks, browsing and purchase data of merchants, pre-process the data to remove duplicate values, segment the data, and obtain the call data S = {ω1, ω2, ..., ω i}, where i = 1, 2, ..., n, ω i For a single word, for each word ω i Embed a vector v i , get the data vector Building a sentiment dictionary Where x = 1, 2, ..., γ, γ is the number of words in the sentiment dictionary, For a single word, for each word Embed a vector ε x , get the sentiment dictionary vector And perform sentiment scoring on each word to get the corresponding sentiment score τ x ;

[0070] Step S222: By calculating the single word vector v in the data vector i With the sentiment dictionary vector ε S Similarity By comparing, the sentiment score of the vector with the smallest similarity is selected as the sentiment score τ of a single data vector i , the sentiment score of the data vector is Call the thresholds μ1 and μ2 set in the database, μ1>μ2. When the sentiment score satisfies τ≤μ2, the sentiment of the data vector is classified as negative sentiment. When the sentiment score satisfies μ1>τ>μ2, the sentiment of the data vector is classified as neutral sentiment. When the sentiment score satisfies μ1≤τ, the sentiment of the data vector is classified as positive sentiment, and the corresponding sentiment score τ is output;

[0071] Step S223: Convert the sentiment score τ result into the characteristic data of consumer behavior, and obtain the characteristic data combined with the sentiment score as f=[r,τ], where r is the characteristic value of the corresponding consumer behavior data. Input the characteristic data into the logistic regression algorithm, and predict the probability of consumers making purchase decisions through the regression logic algorithm. where w T The weight set in the database is used to call the threshold δ1 set in the database. When δ≥δ1, the consumer makes a purchase, and when δ<δ1, the consumer does not make a purchase.

[0072] In an embodiment of the present invention, step S3 further includes the following steps:

[0073] Step S31: call the feature data in the commodity database including commodity category x i 、Brandy i 、Price i , we get the eigenvector N = [x i ,y i ,z i ], when the system determines that the consumer may make a purchase, the system collects data on the goods that the consumer may purchase, pre-processes the data, and extracts features including product category x, brand y, and price z to obtain a feature vector M = [x, y, z];

[0074] Step S32: Calculate the distance between the product feature vector M and the feature vector N of the product in the product database to obtain the similarity Where n is the number of products in the product database, the threshold d0 set in the database is called, and the products in the product database that meet d i Products with value ≥d0 are pushed to consumers.

[0075] In an embodiment of the present invention, step S4 further includes the following steps:

[0076] Step S41: call the user transaction records in the consumer behavior database including transaction time and amount, preprocess the data experience, and extract features including consumption frequency K and consumption amount distribution Q, call the user transaction record features in the historical consumer behavior database including consumption frequency K i 、Consumption amount distribution Q i , normalize the data;

[0077] Step S42: By formula Calculate the degree of deviation α between the user transaction record characteristics and the historical user transaction record characteristics K , α Q , where X is the type of user transaction record feature, is the average value of historical user transaction record data, σ is the standard deviation of historical user transaction record data, and the total deviation degree is obtained by calculation where ω K ,ω Q For the corresponding weight, the threshold α0 set in the database is called. When α≥α0, the system determines that the consumer behavior is abnormal and there is consumer fraud, and the system consumer is warned.

[0078] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0079] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A consumer behavior prediction platform based on big data, characterized by: The consumer behavior prediction platform includes: Data management module, used to manage the data collected by the consumer behavior prediction platform; Behavior prediction module, used to predict and evaluate consumers' purchasing behavior and recommend satisfactory products to consumers; The warning module is used to detect abnormal consumer behavior and warn consumers.

2. The consumer behavior prediction platform based on big data according to claim 1, characterized in that: The data management module comprises: A data collection unit, used to collect consumer behavior data and product information data; The data processing unit is used to pre-process the collected data and extract corresponding features.

3. The consumer behavior prediction platform based on big data according to claim 2 is characterized by: The data management module comprises: A consumption desire analysis unit is used to analyze consumers' consumption desire through consumer behavior data and make predictions, and to make consumer behavior predictions in months when consumers' consumption desire is medium or high; The emotion prediction unit is used to analyze the emotion of consumers by analyzing the consumer behavior data, and predict the consumer behavior by the influence of emotion on the consumer behavior; The product recommendation unit is used to analyze consumers' purchasing behavior and recommend related products to consumers.

4. The consumer behavior prediction platform based on big data according to claim 3 is characterized by: The warning module comprises: Abnormal behavior detection unit, used to analyze consumers’ purchasing behavior and detect whether the behavior is abnormal; The warning unit is used to warn consumers when their behavior is abnormal.

5. A method for operating a consumer behavior prediction platform based on big data, comprising: Step S1: Establish a database, collect consumer behavior data and product information data in real time, store the data in the database, and import historical consumer behavior data into the database; Step S2: calling the consumer behavior data in the database, preprocessing the data, and extracting features, analyzing the feature data to obtain the consumer's consumption desire parameters, and predicting the consumer's consumption behavior by analyzing the consumer's emotions based on the consumption desire parameters; Step S3: Based on the results of consumer behavior prediction, when the system determines that the consumer is likely to make a purchase, the system automatically pushes relevant product information to the consumer; Step S4: While predicting consumer behavior, analyze the possible purchasing behavior of consumers to determine whether the consumer's purchasing behavior is abnormal and issue a warning.

6. The method for operating a consumer behavior prediction platform based on big data according to claim 5, characterized in that: The database is established to collect consumer behavior data in real time, including user transaction records, social media data, and mobile application behavior data. The user transaction records include transaction time, transaction location, amount, and product type. The social media data includes product comments, likes, and sharing data. The mobile application behavior data includes the click, browse, and purchase data of the merchant. After pre-processing the collected data, it is stored in the database, and the historical consumer behavior data is imported into the system; Establish a product database and extract the features of product information data including product category, brand, and price. Different product categories and brands have corresponding codes, and the code does not exceed 1. The price is normalized and the feature data is stored in the product database.

7. The method for operating a consumer behavior prediction platform based on big data according to claim 6, characterized in that: The step S2 further comprises the following steps: Step S21: call the historical consumer behavior data in the database, pre-process the data, and extract the features in the data including the consumption frequency F i 、Average Consumption A i , where i = 1, 2, ..., 12, the consumption amount growth rate is calculated By calculating the consumer's consumption desire parameter D i =ω F F i +ω A A i +ω R R i , where ω F ,ω A ,ω R For the weight data set up in the database, the consumption desire parameters are saved in the database according to time; Step S22: Call the consumption desire parameter data D in the database i , call the set consumption desire parameter thresholds D1, D2 in the database, D1 < D2, when D i ≤ D1, it is determined that the consumption desire of consumers in this month is low, when D2 > D i > D1, it is determined that the consumption desire of consumers in this month is medium, when D2 ≤ D i , it is determined that the consumption desire of consumers in this month is high. Based on the judgment result, predict the consumption behavior of consumers in the next months when the consumption desire of consumers is high and medium.

8. The method for operating a consumer behavior prediction platform based on big data according to claim 7, characterized in that: The step S22 further comprises the following steps: Step S221: call the social media data in the database including comments, likes data and mobile application behavior data including clicks, browsing and purchase data of merchants, pre-process the data to remove duplicate values, segment the data, and obtain the call data S = {ω1, ω2, ..., ω i }, where i = 1, 2, ..., n, ω i For a single word, for each word ω i Embed a vector v i , get the data vector Building a sentiment dictionary Where x = 1, 2, ..., γ, γ is the number of words in the sentiment dictionary, For a single word, for each word Embed a vector ε x , get the sentiment dictionary vector And perform sentiment scoring on each word to get the corresponding sentiment score τ x ; Step S222: By calculating the single word vector v in the data vector i With the sentiment dictionary vector ε S Similarity By comparing, the sentiment score of the vector with the smallest similarity is selected as the sentiment score τ of a single data vector i , the sentiment score of the data vector is Call the thresholds μ1 and μ2 set in the database, μ1>μ2. When the sentiment score satisfies τ≤μ2, the sentiment of the data vector is classified as negative sentiment. When the sentiment score satisfies μ1>τ>μ2, the sentiment of the data vector is classified as neutral sentiment. When the sentiment score satisfies μ1≤τ, the sentiment of the data vector is classified as positive sentiment, and the corresponding sentiment score τ is output; Step S223: Convert the sentiment score τ result into the characteristic data of consumer behavior, and obtain the characteristic data combined with the sentiment score as f=[r,τ], where r is the characteristic value of the corresponding consumer behavior data. Input the characteristic data into the logistic regression algorithm, and predict the probability of consumers making purchase decisions through the regression logic algorithm. where w T The weight set in the database is used to call the threshold δ1 set in the database. When δ≥δ1, the consumer makes a purchase, and when δ<δ1, the consumer does not make a purchase.

9. The method for operating a consumer behavior prediction platform based on big data according to claim 8, characterized in that: The step S3 further comprises the following steps: Step S31: call the feature data in the commodity database including commodity category x i 、Brandy i 、Price i , we get the eigenvector N = [x i ,y i ,z i ], when the system determines that the consumer may make a purchase, the system collects data on the goods that the consumer may purchase, pre-processes the data, and extracts features including product category x, brand y, and price z to obtain a feature vector M = [x, y, z]; Step S32: Calculate the distance between the product feature vector M and the feature vector N of the product in the product database to obtain the similarity Where n is the number of products in the product database, the threshold d0 set in the database is called, and the products in the product database that meet d i Products with value ≥d0 are pushed to consumers.

10. The method for operating a consumer behavior prediction platform based on big data according to claim 9, characterized in that: The step S4 further comprises the following steps: Step S41: call the user transaction records in the consumer behavior database including transaction time and amount, preprocess the data experience, and extract features including consumption frequency K and consumption amount distribution Q, call the user transaction record features in the historical consumer behavior database including consumption frequency K i 、Consumption amount distribution Q i , normalize the data; Step S42: By formula Calculate the degree of deviation α between the user transaction record characteristics and the historical user transaction record characteristics K , α Q , where X is the type of user transaction record feature, is the average value of historical user transaction record data, σ is the standard deviation of historical user transaction record data, and the total deviation degree is obtained by calculation where ω K ,ω Q For the corresponding weight, the threshold α0 set in the database is called. When α≥α0, the system determines that the consumer behavior is abnormal and there is consumer fraud, and the system consumer is warned.

Images