Data analysis method based on big data and related equipment
By analyzing the behavior records of the target users to generate personalized tags, and generating and optimizing push information based on these tags, the problem of difficulty in dealing with and analyzing large-scale data is solved, and precise message push and satisfaction are improved with users.
Patent Information
- Application Number
- CN202510361541.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-26
- Publication Date
- 2025-04-25
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional data processing and analysis methods are difficult to deal with large-scale and complex data, and how to efficiently extract meaningful information from big data and generate user-satisfied push messages has become a key challenge.
By obtaining the user information and historical behavior records of the target user, analyzing and generating user personalized tags, generating personalized push information based on these tags, and adjusting and optimizing push content through feedback information to achieve accurate message push.
This improves the user's satisfaction with the pushed messages, reduces the occurrence of unsatisfaction or irrelevance of the push messages, and realizes accurate message push to the target user.
Smart Images
Figure CN119884490A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data analysis technology, and in particular to a data analysis method based on big data and related equipment. Background Art
[0002] Various devices, sensors, social media platforms, e-commerce websites, etc. are constantly generating and accumulating large amounts of data every day, and the amount of data is showing an explosive growth trend. These data are multi-source, heterogeneous, and high-dimensional. On the one hand, traditional data processing and analysis methods are difficult to cope with such large-scale and complex data. On the other hand, massive amounts of data contain huge value, but how to efficiently extract meaningful information from these data and generate push messages suitable for each user, thereby improving user satisfaction with the pushed messages, has become a key challenge faced by many fields. Summary of the invention
[0003] In order to help improve user satisfaction with pushed messages, the present application provides a data analysis method and related equipment based on big data.
[0004] In the first aspect, the present application provides a data analysis method based on big data, which adopts the following technical solution: A data analysis method based on big data, comprising: Obtain user information corresponding to the target user; Based on the user information, obtaining the historical behavior records of the target user; Based on the historical behavior records, the target user is analyzed, and a user personality tag corresponding to the target user is obtained; Based on the user's personality tag, obtaining personalized push information; Obtaining feedback information from the target user on the pushed information; Based on the feedback information, generate user interest tags; Based on the user interest tags, generating target push information; The step of analyzing the target user based on the historical behavior record and obtaining the user personality tag corresponding to the target user includes: Determine whether the historical behavior record is an empty record; If the historical behavior record is the empty record, obtaining the associated users of the target user; Acquire the associated tag of the associated user, and use the associated tag as the user personality tag corresponding to the target user; If the associated tag is not obtained, determining whether the target user is a new user; If the target user is the new user, obtaining the information popularity ranking; Based on the information popularity ranking, obtaining the user personality tag; If the historical behavior record is not the empty record, obtaining the search history, browsing history and transaction history of the target user based on the historical behavior record; Based on the search history, the browsing history and the transaction history, a user personality tag is obtained.
[0005] By adopting the above technical solution, historical behavior records are obtained according to the user information of the target user, and then the target user is analyzed according to the target behavior records, and the user personality tag corresponding to the target user is obtained. The system generates personalized push information suitable for the target user according to the user personality tag of the target user and pushes it to the target user, and then collects the feedback information of the target user on the personalized push information. The system then marks the push messages that the target user is really interested in with user interest tags according to the feedback information, and finally pushes the target push information marked with the user interest tag to the target user; On the basis of generating personalized push information according to user personality tags, target push information that users are satisfied with and truly interested in is selected based on feedback information. This helps reduce situations where target users are dissatisfied with personalized push information generated based on historical behavior records. It also helps reduce situations where the system continues to push personalized push information to target users that they were interested in in the past but are not interested in now. This helps to achieve accurate message push to target users, and further helps to improve target users' satisfaction with the pushed messages.
[0006] Optionally, obtaining the associated tag of the associated user and using the associated tag as the user personality tag corresponding to the target user includes: Acquire the degree of association between the target user and the associated user; Determining whether the correlation degree exceeds a preset correlation threshold; If the association degree does not exceed the preset association threshold, determining whether the associated user has a privacy setting; If the associated user has the privacy setting, acquiring a visible tag based on the associated tag and the privacy setting; Using the visual tag as a user personality tag corresponding to the target user; If the associated user does not have the privacy setting, the associated tag is used as the user personality tag corresponding to the target user.
[0007] Optionally, the correlation degree includes full correlation, strong correlation and weak correlation; and obtaining the correlation degree between the target user and the associated user includes: Acquire a user relationship between the target user and the associated user; If the target user and the associated user are common users, the full association is used as the association degree between the target user and the associated user; If the target user and the associated user are not the common users, obtaining the intimacy between the target user and the associated user; If the intimacy exceeds a preset intimacy threshold, the strong association is used as the association between the target user and the associated user; If the intimacy does not exceed a preset intimacy threshold, the weak association is used as the association between the target user and the associated user.
[0008] Optionally, if the target user and the associated user are not common users, obtaining the intimacy between the target user and the associated user includes: If the target user and the associated user are not the common users, obtaining the communication records and transaction records between the target user and the associated user; Based on the communication records, obtaining the number of communications and the content of communications; Analyze the communication content and obtain a comprehensive score corresponding to the communication content; Based on the transaction records, obtain the transaction times and transaction amounts corresponding to different exchanges; Based on the number of communications, the comprehensive score, the number of transactions and the transaction amount, obtaining the intimacy between the target user and the associated user; The intimacy satisfies the following calculation formula: Among them, I is the intimacy, represents the weight of the j-th indicator, For different indicators The calculation function of .
[0009] Optionally, analyzing the communication content and obtaining a comprehensive score corresponding to the communication content includes: Preprocessing the communication content and obtaining the preprocessed target content; Extracting features from the target content and generating target features; The target features are input into a preset scoring model, and a comprehensive score is output.
[0010] Optionally, the specific steps include: The generating of user interest tags based on the feedback information includes: Based on the feedback information, obtaining the target user's browsing history of the personalized push information; Based on the browsing history, obtain the number of views, initial browsing duration, single browsing duration change, and total browsing duration of different personalized push information; Based on the number of browsing times, the initial browsing duration, the change in the single browsing duration, and the total browsing duration, obtaining the interest value of the target user corresponding to different personalized push information; Determining whether the interest value exceeds a tag generation threshold; If the interest value exceeds the tag generation threshold, generating an interest tag based on the interest value; The interest value satisfies the following calculation formula: Among them, S is the interest value, N is the number of views, T1 is the first view duration, Tt is the total view duration, Ti is the i-th view duration, ΔT is the change in single view duration, is the weight of the number of views, is the weight of the first browsing duration, is the weight of the total browsing time, is the weight of a single duration change.
[0011] In the second aspect, the present application also discloses a data analysis system based on big data, which adopts the following technical solution: A data analysis system based on big data, comprising: A first acquisition module is used to acquire user information corresponding to a target user; A second acquisition module, used to acquire the historical behavior record of the target user based on the user information; A third acquisition module is used to analyze the target user based on the historical behavior record and obtain a user personality tag corresponding to the target user; A fourth acquisition module, used to acquire personalized push information based on the user's personality tag; A fifth acquisition module, used to acquire feedback information of the target user on the pushed information; A first generating module, used to generate user interest tags based on the feedback information; The second generating module is used to generate target push information based on the user interest tag.
[0012] By adopting the above technical solution, historical behavior records are obtained according to the user information of the target user, and then the target user is analyzed according to the target behavior records, and the user personality tag corresponding to the target user is obtained. The system generates personalized push information suitable for the target user according to the user personality tag of the target user and pushes it to the target user, and then collects the feedback information of the target user on the personalized push information. The system then marks the push messages that the target user is really interested in with user interest tags according to the feedback information, and finally pushes the target push information marked with the user interest tag to the target user; On the basis of generating personalized push information according to user personality tags, target push information that users are satisfied with and truly interested in is selected based on feedback information. This helps reduce situations where target users are dissatisfied with personalized push information generated based on historical behavior records. It also helps reduce situations where the system continues to push personalized push information to target users that they were interested in in the past but are not interested in now. This helps to achieve accurate message push to target users, and further helps to improve target users' satisfaction with the pushed messages.
[0013] In a third aspect, the present application provides a computer device, which adopts the following technical solution: An intelligent terminal comprises a memory and a processor, wherein the memory is used to store a computer program that can be run on the processor, and when the processor loads the computer program, the method of the first aspect is executed.
[0014] By adopting the above technical solution, a computer program is generated based on the method of the first aspect and stored in a memory to be loaded and executed by a processor, thereby manufacturing an intelligent terminal based on the memory and the processor for user convenience.
[0015] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium stores a computer program, and when the computer program is loaded by a processor, the method of the first aspect is executed.
[0016] By adopting the above technical solution, a computer program is generated based on the method of the first aspect and stored in a computer-readable storage medium so as to be loaded and executed by a processor. The computer-readable storage medium facilitates the readability and storage of the computer program.
[0017] In summary, this application includes the following beneficial technical effects: On the basis of generating personalized push information according to user personality tags, target push information that users are satisfied with and truly interested in is selected based on feedback information. This helps reduce situations where target users are dissatisfied with personalized push information generated based on historical behavior records. It also helps reduce situations where the system continues to push personalized push information to target users that they were interested in in the past but are not interested in now. This helps to achieve accurate message push to target users, and further helps to improve target users' satisfaction with the pushed messages. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Figure 1 This is a main flow chart of a data analysis method based on big data in an embodiment of the present application; Figure 2 is a flowchart of steps S201 to S208; Figure 3 is a flowchart of steps S301 to S306; Figure 4 is a flowchart of steps S401 to S405; Figure 5 is a flowchart of steps S501 to S505; Figure 6 is a flowchart of steps S601 to S603; Figure 7 is a flowchart of steps S701 to S705; Figure 8 It is a module diagram of a data analysis system based on big data in an embodiment of the present application.
[0019] Description of reference numerals: 1. First acquisition module; 2. Second acquisition module; 3. Third acquisition module; 4. Fourth acquisition module; 5. Fifth acquisition module; 6. First generation module; 7. Second generation module. DETAILED DESCRIPTION
[0020] In a first aspect, the present application discloses a data analysis method based on big data.
[0021] Reference Figure 1 , a data analysis method based on big data, comprising steps S101 to S107: Step S101: Obtain user information corresponding to the target user.
[0022] Specifically, in this embodiment, the target user is the user who needs to perform data analysis, and the user information is all relevant information corresponding to the target user, including the user's account information, login information, and operation information.
[0023] Step S102: Based on the user information, the historical behavior records of the target user are obtained.
[0024] Specifically, in this embodiment, the historical behavior records are all behavior records of the account corresponding to the target user, including search records, browsing records, and transaction records.
[0025] Step S103: Analyze the target user based on the historical behavior records, and obtain the user personality label corresponding to the target user.
[0026] Specifically, in this embodiment, the historical behavior records of the target user are analyzed to obtain the information related to the target user, and the information is marked, and the mark is the user personality label.
[0027] Step S104: Obtain personalized push information based on the user's personality tag.
[0028] Specifically, in this embodiment, the personalized push information is push information generated according to the user's personalized tag.
[0029] Step S105: Obtaining feedback information of the target user on the pushed information.
[0030] Specifically, in this embodiment, after the target user receives the personalized push information pushed by the system, if the target user browses the personalized recommendation message or generates related transaction behavior, corresponding records and information will be generated, and the records and information are feedback information for the push information.
[0031] Step S106: Generate user interest tags based on the feedback information.
[0032] Specifically, in this embodiment, it is possible to know which of the personalized push information the target user is satisfied with or truly interested in based on the feedback information, and generate a corresponding tag for this type of push information, which is the user interest tag.
[0033] Step S107: Generate target push information based on user interest tags.
[0034] Specifically, in this embodiment, according to the user interest tags, the push information that the target user is truly satisfied with and interested in can be obtained, and this type of information is the target push information.
[0035] The data analysis method based on big data provided in this embodiment obtains historical behavior records based on the user information of the target user, then analyzes the target user based on the target behavior records, and obtains the user personality tag corresponding to the target user. The system generates personalized push information suitable for the target user based on the user personality tag of the target user and pushes it to the target user, then collects the target user's feedback information on the personalized push information, and the system then marks the push messages that the target user is really interested in with user interest tags based on the feedback information, and finally pushes the target push information marked with the user interest tags to the target user.
[0036] On the basis of generating personalized push information according to user personality tags, target push information that users are satisfied with and truly interested in is selected based on feedback information. This helps reduce situations where target users are dissatisfied with personalized push information generated based on historical behavior records. It also helps reduce situations where the system continues to push personalized push information to target users that they were interested in in the past but are not interested in now. This helps to achieve accurate message push to target users, and further helps to improve target users' satisfaction with the pushed messages.
[0037] Reference Figure 2 In one implementation of this embodiment, step S103 analyzes the target user based on the historical behavior record and obtains the user personality tag corresponding to the target user, including steps S201 to S208: Step S201: Determine whether the historical behavior record is an empty record.
[0038] Specifically, an empty record means that there is no operation record, which means that after the target user registered an account, he or she has not browsed any information, searched for any information, or performed any transactions.
[0039] Step S202: If the historical behavior record is an empty record, the associated users of the target user are obtained.
[0040] Specifically, associated users are other users associated with the target user, such as recommending users, friend users or followed users. In this embodiment, if the target user is registered after being recommended by other users, such as by clicking on the recommendation link of other users or scanning the recommendation code of other users, then the user is the recommended user of the target user, friend users are other users who have mutually applied to be friends with the target user, and followed users are users followed by the target user.
[0041] Step S203: Acquire the associated tag of the associated user, and use the associated tag as the user personality tag corresponding to the target user.
[0042] Specifically, in this embodiment, the associated tag is the user interest tag corresponding to the associated user. When the historical behavior record of the target user is empty and the associated tag of the associated user corresponding to the target user is obtained, the associated tag is directly used as the user personality tag of the target user.
[0043] Step S204: If no associated tag is obtained, it is determined whether the target user is a new user.
[0044] Specifically, a new user is a user with a short registration time. In this embodiment, a user with a registration time of no more than ten minutes may be regarded as a new user, or it may be set according to user needs.
[0045] Step S205: If the target user is a new user, obtain the information popularity ranking.
[0046] Step S206: Based on the information popularity ranking, obtain the user's personality label.
[0047] Specifically, in this embodiment, if the target user is a new user, it means that the target user may not have performed relevant behavioral operations, and the system cannot accurately form a user personality label for the target user. Therefore, the information popularity ranking list of the current time node can be obtained, and from the information popularity ranking list, several categories of information with high popularity are selected and regarded as information associated with the target user, and a corresponding user personality label is generated.
[0048] Step S207: If the historical behavior record is not an empty record, then based on the historical behavior record, the search record, browsing record and transaction record of the target user are obtained.
[0049] Specifically, in this embodiment, if the historical behavior record is not an empty record, it means that the target user has performed at least one of the following behaviors: search, browse, or transaction, and the corresponding relevant records can be obtained. If only one or two of the behaviors are performed, the records corresponding to the behaviors that have not been performed are empty records.
[0050] Step S208: Obtain user personality tags based on search records, browsing records and transaction records.
[0051] The data analysis method based on big data provided in this embodiment first determines whether the historical behavior record is an empty record. If so, it means that the target user has not performed any search, browsing, transaction and other behaviors after the account registration. The associated users of the target user are further obtained, and the associated tags are used as the user personality tags of the target user. If the associated tags are not obtained, it is determined whether the target user is a new user. If so, it means that the target user's account has been registered not long ago, so no search, browsing and transaction behaviors have been performed. The system cannot generate a user personality tag related to the target user based on this. Therefore, several types of information with high popularity are selected from the information popularity ranking, which are regarded as information associated with the target user, and the corresponding user personality tags are generated.
[0052] If the historical behavior record is not an empty record, there must be corresponding operation traces. Therefore, the corresponding user personality label can be generated based on the search record, browsing record and transaction record of the target user with relevant data.
[0053] Through multiple judgments, the corresponding method is selected according to different situations to obtain the user's personal label, which helps to make the relationship between the user's personal label and the target user closer, thereby helping to improve the target user's satisfaction with the push message.
[0054] Reference Figure 3 In one implementation of this embodiment, step S203 obtains the associated tag of the associated user, and uses the associated tag as the user personality tag corresponding to the target user, including steps S301 to S306: Step S301: Obtain the degree of association between the target user and the associated users.
[0055] Specifically, in this embodiment, the correlation degree is data representing the degree of correlation between the target user and the associated users. The higher the correlation degree, the closer the correlation between the target user and the associated users.
[0056] Step S302: Determine whether the correlation degree exceeds a preset correlation threshold.
[0057] Specifically, the preset correlation threshold is a pre-set judgment standard for judging whether the correlation degree meets the relevant requirements.
[0058] Step S303: If the association degree does not exceed the preset association threshold, it is determined whether the associated user has any privacy settings.
[0059] Specifically, if the correlation degree does not exceed the preset correlation threshold, it means that the correlation degree is not particularly high. When the associated user has certain privacy settings, the target user cannot obtain all the user personality tags of the associated user. Therefore, it is further determined whether the associated user has privacy settings. In this embodiment, the privacy settings are used to distinguish between ordinary user personality tags and more private user personality tags. Only when the correlation degree exceeds the preset correlation threshold, the target user can obtain the privacy settings of the associated user and the user personality tags set to private.
[0060] Step S304: If the associated user has privacy settings, a visible tag is obtained based on the associated tag and the privacy settings.
[0061] Specifically, in this embodiment, the visible tags are all user personal tags corresponding to the associated users except for the user personal tags set as private tags.
[0062] Step S305: Using the visible tag as the user personality tag corresponding to the target user.
[0063] Step S306: If the associated user does not have any privacy settings, the associated tag is used as the user personality tag corresponding to the target user.
[0064] The data analysis method based on big data provided in this embodiment obtains the correlation between the target user and the associated users, and determines whether the correlation exceeds a preset correlation threshold. If not, it means that the correlation is low. When the associated user has user personality tags with high privacy, the target user cannot obtain all of them. Therefore, it is further determined whether the associated user has privacy settings. If so, the target user can only obtain the visual tags of the associated user except the privacy settings as the corresponding user personality tags; if not, all associated tags are used as the user personality tags corresponding to the target user; the correlation is set to determine whether the target user can obtain all the user personality tags of the associated users, and the user personality tags of the target user are determined by determining whether the associated user has privacy settings, which helps the system to more accurately generate the user personality tags corresponding to the target user.
[0065] Reference Figure 4 In one implementation of this embodiment, step S301 of obtaining the degree of association between the target user and the associated user includes steps S401 to S405: Step S401: Obtain the user relationship between the target user and the associated users.
[0066] Specifically, in this embodiment, the user relationship includes shared users and non-shared users, wherein shared users indicate that the target user and the associated user are owned by the same user, and non-shared users indicate that the target user and the associated user are owned by different users.
[0067] Step S402: If the target user and the associated user are common users, the full association is used as the association degree between the target user and the associated user.
[0068] Specifically, the degree of association includes full association, strong association and weak association, wherein full association indicates that the target user and the associated user are fully associated, strong association indicates that the target user and the associated user have a strong association, and weak association indicates that the target user and the associated user have a weak association. In this embodiment, full association can be represented by an association degree of 100%, strong association can be represented by an association degree greater than or equal to 80% and less than 100%, and weak association can be represented by an association degree less than 80%.
[0069] Step S403: If the target user and the associated user are not common users, the intimacy between the target user and the associated user is obtained.
[0070] Specifically, in this embodiment, intimacy refers to the degree of intimacy.
[0071] Step S404: If the intimacy exceeds a preset intimacy threshold, a strong association is taken as the association between the target user and the associated user.
[0072] Specifically, in this embodiment, the preset intimacy threshold is a pre-set judgment standard for judging whether the intimacy reaches a strong correlation.
[0073] Step S405: If the intimacy does not exceed the preset intimacy threshold, the weak association is taken as the association between the target user and the associated user.
[0074] The data analysis method based on big data provided in this embodiment first obtains the user relationship between the target user and the associated user. If the user relationship is a shared user, it means that the target user and the associated user are both owned by the same user. Therefore, it is directly determined that the correlation between the target user and the associated user is fully correlated. If they are not shared users, it means that the target user and the associated user are not owned by the same person. In order to determine the correlation between the two, the intimacy between the two is further obtained, and it is determined whether the intimacy exceeds a preset intimacy threshold. If it exceeds, it means that the intimacy between the two is high. Therefore, the correlation between the target user and the associated user is strongly correlated. If it does not exceed, it means that the intimacy between the two is low. Therefore, it is determined that the correlation between the target user and the associated user is weakly correlated. The correlation between the target user and the associated user is determined based on the user relationship and intimacy, which helps the system to more accurately generate a user personality label corresponding to the target user.
[0075] Reference Figure 5 In one implementation of this embodiment, if the target user and the associated user are not common users in step S403, obtaining the intimacy between the target user and the associated user includes steps S501 to S505: Step S501: If the target user and the associated user are not common users, then the communication records and transaction records between the target user and the associated user are obtained.
[0076] Specifically, in this embodiment, the communication records are all records of communication between the target user and the associated users, including chat records and interaction records, etc.; the transaction records are all records of transactions between the target user and the associated users, including purchase records and sales records, etc.
[0077] Step S502: Based on the communication records, obtain the communication times and communication contents.
[0078] Specifically, in this embodiment, the number of communications refers to the number of times the two parties communicate, and the content of communication refers to the content of each communication between the two parties.
[0079] Step S503: Analyze the communication content and obtain the comprehensive score corresponding to the communication content.
[0080] Specifically, the communication content is analyzed first, and then a corresponding comprehensive score is obtained according to the communication content. In this embodiment, the comprehensive score is one of the criteria for representing the intimacy between the target user and the associated users.
[0081] Step S504: Based on the transaction records, the transaction times and the transaction amounts corresponding to different exchanges are obtained.
[0082] Specifically, in this embodiment, the number of transactions refers to the total number of transactions between the target user and the associated users, and the transaction amount refers to the amount corresponding to each transaction.
[0083] Step S505: Based on the number of communications, the comprehensive score, the number of transactions and the transaction amount, the intimacy between the target user and the associated users is obtained.
[0084] Specifically, in this embodiment, the intimacy satisfies the following calculation formula: Among them, I is the intimacy, represents the weight of the j-th indicator, For different indicators The calculation function of .
[0085] The calculation function of the number of communications x1 satisfies the following calculation formula: Among them, d1 is the attenuation coefficient, which indicates the degree to which the rate of intimacy improvement gradually slows down with the increase in the number of communications; s1 is the saturation coefficient, which indicates the threshold at which the intimacy improvement is no longer significant after the number of communications reaches a certain level.
[0086] After the number of communications reaches a certain level, further communications will have limited effect on improving intimacy. This can be set based on actual conditions and user needs. For example, if past experience shows that when the number of communications reaches about 50, further communications will not significantly improve intimacy, you can try to set d1 to a value that reflects this change. First set s1 = 50, then perform a simple test and adjust d1, for example, first set d1 = 0.02, and substitute it into the formula to see the calculated value. As x1 increases, fine-tuning is performed to see if the changing trend is in line with expectations.
[0087] The calculation function of the comprehensive score x2 satisfies the following calculation formula: Among them, p represents the number of times the chat content has pleasant features, q represents the number of times the chat content has no obvious emotional expression, and r represents the number of times the chat content has unpleasant features. is the score value corresponding to a chat content with pleasant features, This is the score corresponding to a chat when there is no obvious emotional expression. is the score value corresponding to the unpleasant feature of a chat content. In this embodiment, Can be set to 2 points. Can be set to 1 point. Can be set to -1 point.
[0088] The calculation function of the number of transactions x3 satisfies the following calculation formula: Wherein, c is a growth index, and in this embodiment, 0<c is less than or equal to 1.
[0089] The calculation function of the transaction amount x4 satisfies the following calculation formula: Among them, g is the average amount of a single transaction.
[0090] The data analysis method based on big data provided in this embodiment successively obtains the number of communications, comprehensive scores, number of transactions and transaction amounts between the target user and the related users, and then calculates the intimacy between the target user and the related users by combining the number of communications, comprehensive scores, number of transactions and transaction amounts. By combining the number of communications, comprehensive scores, number of transactions and transaction amounts at the same time, it helps to calculate the intimacy more accurately.
[0091] Reference Figure 6 In one implementation of this embodiment, step S503 analyzes the communication content and obtains the comprehensive score corresponding to the communication content, including steps S601 to S603: Step S601: pre-process the communication content and obtain the pre-processed communication content.
[0092] Specifically, in this embodiment, the preprocessing includes text cleaning, removing irrelevant characters, punctuation marks, special characters, etc., and unifying the text format; the communication content is the communication content after preprocessing.
[0093] Step S602: extract features from the communication content and generate target features.
[0094] Specifically, in this embodiment, the target feature is a feature that characterizes the emotion of the chat content, such as a feature that the chat content is pleasant, a feature that the chat content is unpleasant, and a feature that the chat content has no obvious emotional expression.
[0095] Step S603: Input the target features into a preset scoring model and output a comprehensive score.
[0096] The data analysis method based on big data provided in this embodiment extracts features of communication content and generates target features, and then outputs a comprehensive score based on the target features. Outputting a comprehensive score based on different target features helps to calculate the comprehensive score more accurately.
[0097] Reference Figure 7 In one implementation of this embodiment, step S106 generates user interest tags based on feedback information, including steps S701 to S705: Step S701: Based on the feedback information, the browsing history of the target user on the personalized push information is obtained.
[0098] Specifically, in this embodiment, the browsing record is the browsing record of the target user on the personalized push information, including the number of browsing times, the initial browsing time, the change in single browsing time, and the total browsing time.
[0099] Step S702: Based on the browsing history, the number of views, initial browsing duration, single browsing duration change, and total browsing duration of different personalized push information are obtained.
[0100] Specifically, in this embodiment, the number of views is the total number of views of the target user on a single personalized push information, the first viewing duration is the duration of the target user's first viewing of the single personalized push information, the single duration change is the change in the duration of the target user's browsing of the single personalized push information, and the total browsing duration is the total duration of the target user's browsing of the single personalized push information.
[0101] Step S703: Based on the number of views, initial view duration, single view duration change, and total view duration, the interest value of the target user corresponding to different personalized push information is obtained.
[0102] Specifically, in this embodiment, the interest value satisfies the following calculation formula: Among them, S is the interest value, N is the number of views, T1 is the first view duration, Tt is the total view duration, Ti is the i-th view duration, ΔT is the change in single view duration, is the weight of the number of views, is the weight of the first browsing duration, is the weight of the total browsing time, is the weight of a single duration change.
[0103] Step S704: Determine whether the interest value exceeds the tag generation threshold.
[0104] Specifically, in this embodiment, the tag generation threshold is the minimum value for generating user interest tags.
[0105] Step S705: If the interest value exceeds the tag generation threshold, an interest tag is generated based on the interest value.
[0106] The data analysis method based on big data provided in this embodiment obtains the browsing history of the target user for the personalized push information according to the feedback information, and then obtains the browsing times, the initial browsing time, the change of the single browsing time and the total browsing time of different personalized push information according to the browsing history; finally, according to the browsing times, the initial browsing time, the change of the single browsing time and the total browsing time, obtains the interest value of the target user for the different personalized push information, and determines whether the interest value exceeds the tag generation threshold. If so, generates an interest tag according to the interest value; calculating the interest value of the target user for the different personalized push information through information of different dimensions helps to improve the accuracy of the interest value.
[0107] On the second aspect, the present application also discloses a data analysis system based on big data.
[0108] Reference Figure 8 , a data analysis system based on big data, comprising: A first acquisition module is used to acquire user information corresponding to a target user; The second acquisition module is used to acquire the historical behavior records of the target user based on the user information; The third acquisition module is used to analyze the target user based on the historical behavior records and obtain the user personality label corresponding to the target user; The fourth acquisition module is used to obtain personalized push information based on the user's personality tag; The fifth acquisition module is used to obtain the target user's feedback information on the pushed information; A first generating module, used to generate user interest tags based on feedback information; The second generating module is used to generate target push information based on user interest tags.
[0109] In a third aspect, an embodiment of the present application discloses an intelligent terminal, including a memory and a processor, wherein the memory is used to store a computer program that can be run on the processor, and when the processor loads the computer program, it executes a data analysis method based on big data of the above embodiment.
[0110] In a fourth aspect, an embodiment of the present application discloses a computer-readable storage medium, and a computer program is stored in the computer-readable storage medium, wherein when the computer program is loaded by a processor, a data analysis method based on big data of the above embodiment is executed.
[0111] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.
Claims
1. A data analysis method based on big data, characterized in that: include: Obtain user information corresponding to the target user; Based on the user information, obtaining the historical behavior records of the target user; Based on the historical behavior records, the target user is analyzed, and a user personality tag corresponding to the target user is obtained; Based on the user's personality tag, obtaining personalized push information; Obtaining feedback information from the target user on the pushed information; Based on the feedback information, generate user interest tags; Generate target push information based on the user interest tag; The step of analyzing the target user based on the historical behavior record and obtaining the user personality tag corresponding to the target user includes: Determine whether the historical behavior record is an empty record; If the historical behavior record is the empty record, obtaining the associated users of the target user; Acquire the associated tag of the associated user, and use the associated tag as the user personality tag corresponding to the target user; If the associated tag is not obtained, determining whether the target user is a new user; If the target user is the new user, obtaining the information popularity ranking; Based on the information popularity ranking, obtaining the user personality tag; If the historical behavior record is not the empty record, obtaining the search history, browsing history and transaction history of the target user based on the historical behavior record; Based on the search history, the browsing history and the transaction history, a user personality tag is obtained.
2. A data analysis method based on big data according to claim 1, characterized in that: The step of obtaining the associated tag of the associated user and using the associated tag as the user personality tag corresponding to the target user includes: Acquire the degree of association between the target user and the associated user; Determining whether the correlation degree exceeds a preset correlation threshold; If the association degree does not exceed the preset association threshold, determining whether the associated user has a privacy setting; If the associated user has the privacy setting, acquiring a visible tag based on the associated tag and the privacy setting; Using the visual tag as a user personality tag corresponding to the target user; If the associated user does not have the privacy setting, the associated tag is used as the user personality tag corresponding to the target user.
3. A data analysis method based on big data according to claim 2, characterized in that: The degree of association includes full association, strong association and weak association; and obtaining the degree of association between the target user and the associated user includes: Acquire a user relationship between the target user and the associated user; If the target user and the associated user are common users, the full association is used as the association degree between the target user and the associated user; If the target user and the associated user are not the common users, obtaining the intimacy between the target user and the associated user; If the intimacy exceeds a preset intimacy threshold, the strong association is used as the association between the target user and the associated user; If the intimacy does not exceed a preset intimacy threshold, the weak association is used as the association between the target user and the associated user.
4. The data analysis method based on big data according to claim 3 is characterized in that: If the target user and the associated user are not common users, obtaining the intimacy between the target user and the associated user includes: If the target user and the associated user are not the common users, obtaining the communication records and transaction records between the target user and the associated user; Based on the communication records, obtaining the number of communications and the content of communications; Analyze the communication content and obtain a comprehensive score corresponding to the communication content; Based on the transaction records, obtain the transaction times and transaction amounts corresponding to different exchanges; Based on the number of communications, the comprehensive score, the number of transactions and the transaction amount, obtaining the intimacy between the target user and the associated user; The intimacy satisfies the following calculation formula: Among them, I is the intimacy, represents the weight of the j-th indicator, For different indicators The calculation function of .
5. The data analysis method based on big data according to claim 4, characterized in that: The analyzing the communication content and obtaining the comprehensive score corresponding to the communication content includes: Preprocessing the communication content and obtaining the preprocessed target content; Extracting features from the target content and generating target features; The target features are input into a preset scoring model, and a comprehensive score is output.
6. The data analysis method based on big data according to claim 1, characterized in that: The generating of user interest tags based on the feedback information includes: Based on the feedback information, obtaining the target user's browsing history of the personalized push information; Based on the browsing history, obtain the number of views, initial browsing duration, single browsing duration change, and total browsing duration of different personalized push information; Based on the number of browsing times, the initial browsing duration, the change in the single browsing duration, and the total browsing duration, obtaining the interest value of the target user corresponding to different personalized push information; Determining whether the interest value exceeds a tag generation threshold; If the interest value exceeds the tag generation threshold, generating an interest tag based on the interest value; The interest value satisfies the following calculation formula: Among them, S is the interest value, N is the number of views, T1 is the first view duration, Tt is the total view duration, Ti is the i-th view duration, ΔT is the change in single view duration, is the weight of the number of views, is the weight of the first browsing duration, is the weight of the total browsing time, is the weight of a single duration change.
7. A data analysis system based on big data, used to execute the method according to any one of claims 1 to 6, characterized in that: include: A first acquisition module is used to acquire user information corresponding to a target user; A second acquisition module, used to acquire the historical behavior record of the target user based on the user information; A third acquisition module is used to analyze the target user based on the historical behavior record and obtain a user personality tag corresponding to the target user; A fourth acquisition module, used to acquire personalized push information based on the user's personality tag; A fifth acquisition module, used to acquire feedback information of the target user on the pushed information; A first generating module, used to generate user interest tags based on the feedback information; The second generating module is used to generate target push information based on the user interest tag.
8. An intelligent terminal, comprising a memory and a processor, characterized in that: The memory is used to store a computer program that can be run on the processor, and when the processor loads the computer program, it executes the method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored therein, characterized in that: When the computer program is loaded by a processor, the method according to any one of claims 1 to 6 is executed.