An intelligent operation method and system based on data mining and analysis

Abstract

The present invention relates to an intelligent operation method and system based on data mining and analysis, and relates to the technical field of data analysis; the method mainly comprises: sending operation information to N first users and M second users to obtain response data of the N first users and the M second users to the operation information respectively; determining the first user demand information and the second user demand information interested in the operation information according to the response data of the N first users and the M second users to the operation information respectively; determining the final user demand information according to the initial user demand information, the first user demand information and the second user demand information; sending the operation information to the corresponding user according to the final user demand information. The present invention can better understand the operation status of each product and business line through the final user demand information, and provide support for the optimization and improvement of decision-making.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to an intelligent operation method and system based on data mining and analysis. Background Art

[0002] Intelligent operation is a way to optimize and improve the business operation process by using advanced information technology, data analysis, artificial intelligence and other means. It uses automation tools and AI technology, focuses on user needs, provides personalized services or products covering the entire life cycle of users, and realizes self-conversion and self-growth of users.

[0003] Most operational activities are still conducted in a broad manner without fully considering the individual differences of consumers. At the same time, traditional operational methods are usually one-way, that is, there is no real interaction between enterprises and consumers, and many enterprises do not fully utilize data analysis and market insights in their marketing operations. That is, the accuracy of pushing relevant information to consumers is low, resulting in poor operational results. Summary of the invention

[0004] The present invention aims to provide an intelligent operation method and system based on data mining and analysis to solve the deficiencies in the prior art. The technical problem to be solved by the present invention is achieved through the following technical solutions.

[0005] An embodiment of the present invention provides an intelligent operation method based on data mining and analysis, the method comprising:

[0006] Obtaining operation information and its corresponding initial user demand information, wherein the operation information includes sales product information, sales time, sales channel information, operation cost, operation profit, and project goals;

[0007] Randomly obtain N first users corresponding to the initial user demand information from the user collection database, and randomly select M second users from the user collection database;

[0008] Sending the operation information to the N first users and the M second users to obtain response data of the N first users and the M second users to the operation information, respectively, the response data including: page click behavior, page browsing time, reservation purchase behavior, interested click behavior, and uninterested click behavior;

[0009] Determining first user demand information and second user demand information interested in the operation information according to the response data of the N first users and the M second users to the operation information respectively;

[0010] Determining final user demand information according to the initial user demand information, the first user demand information and the second user demand information;

[0011] The operation information is sent to the corresponding user according to the final user demand information, so that the user's terminal device displays the operation information.

[0012] In an optional embodiment, the randomly acquiring N first users corresponding to the initial user demand information from a user collection library, and randomly extracting M second users from the user collection library, includes:

[0013] Determine target users corresponding to the initial user demand information in the user set database, and randomly select N first users from the target users;

[0014] M second users are randomly selected from the user set library except the target user, where the number of M is greater than N.

[0015] In an optional embodiment, determining the first user demand information and the second user demand information interested in the operation information according to the response data of the N first users and the M second users to the operation information respectively includes:

[0016] Performing clustering calculation on the response data of the N first users to the operation information to obtain a first cluster; performing clustering calculation on the response data of the M second users to the operation information to obtain a second cluster;

[0017] Determine first user demand information interested in the operation information according to the cluster center of the first cluster; and determine second user demand information interested in the operation information according to the cluster center of the second cluster.

[0018] In an optional embodiment, determining final user demand information according to the initial user demand information, the first user demand information and the second user demand information includes:

[0019] Performing clustering calculations on each field in the initial user demand information, the first user demand information, and the second user demand information to obtain a field clustering result;

[0020] The corresponding user field portrait is determined according to the cluster center in the field clustering result of each field; and the user field portraits of all fields are combined into the final user demand information.

[0021] In an optional embodiment, determining final user demand information according to the initial user demand information, the first user demand information and the second user demand information includes:

[0022] Performing clustering calculation on the first user demand information and the second user demand information, and determining a cluster center in a clustering result obtained by the clustering calculation as the intermediate user demand information;

[0023] The final user demand information is obtained by intersecting the fields of the intermediate user demand information and the initial user demand information.

[0024] In an optional embodiment, after determining the first user demand information and the second user demand information that are interested in the operation information, the method further includes:

[0025] If the number of first users and second users who are interested in the operation information is less than the corresponding preset value, the operation information is updated according to the first user demand information and the second user demand information who are interested in the operation information, and the step of sending the operation information to the N first users and the M second users is jumped to continue the execution until the number of first user demand information and second user demand information who are interested in the operation information is greater than or equal to the corresponding preset value.

[0026] In an optional embodiment, updating the operation information according to the first user demand information and the second user demand information of interest in the operation information includes:

[0027] Obtaining interest preference information corresponding to the first user demand information and the second user demand information of interest to the operation information;

[0028] Converting the interest preference information into a feature vector matrix;

[0029] Inputting the characteristic vector matrix into the operation information prediction model to obtain corresponding operation prediction information, wherein the operation prediction information includes predicted sales product information, sales time, sales channel information, operation cost, operation profit, and project goal;

[0030] The operation information is updated according to the operation forecast information.

[0031] In an optional embodiment, converting the interest preference information into a feature vector matrix includes:

[0032] Determining a weight value of each interest preference information based on the number of times the same preference information appears in the interest preference information;

[0033] Determine the interest preference information whose weight value is greater than the target weight value as the target interest preference information;

[0034] The target interest preference information and its corresponding weight value are converted into the feature vector matrix.

[0035] In an optional embodiment, the training process of the operation information prediction model is:

[0036] Obtaining sample user interest preference information and corresponding operational annotation information, wherein the operational annotation information includes predicted sales product information, sales time, and sales channel information;

[0037] Converting the interest preference information of the sample users into a sample feature vector matrix;

[0038] Inputting the sample feature vector matrix into the operation information prediction model to obtain corresponding operation prediction information;

[0039] Calculating a loss value according to the operation prediction information and the operation annotation information;

[0040] When the loss value is less than a preset loss value, the training of the operation information prediction model is stopped.

[0041] An embodiment of the present invention provides an intelligent operation system based on data mining and analysis, and the intelligent operation system based on data mining and analysis includes:

[0042] An acquisition module is used to acquire operation information and its corresponding initial user demand information, wherein the operation information includes sales product information, sales time, sales channel information, operation cost, operation profit, and project goal;

[0043] The acquisition module is further used to randomly acquire N first users corresponding to the initial user demand information from the user collection library, and randomly select M second users from the user collection library;

[0044] A sending module, used for sending the operation information to the N first users and the M second users, so as to obtain response data of the N first users and the M second users to the operation information respectively, wherein the response data includes: page click behavior, page browsing time, reservation purchase behavior, interested click behavior, and uninterested click behavior;

[0045] A determination module, configured to determine first user demand information and second user demand information interested in the operation information according to response data of the N first users and the M second users to the operation information respectively;

[0046] The determining module is further configured to determine final user demand information based on the initial user demand information, the first user demand information and the second user demand information;

[0047] The sending module is further used to send the operation information to the corresponding user according to the final user demand information, so that the user's terminal device displays the operation information.

[0048] The embodiments of the present invention include the following advantages:

[0049] An intelligent operation method and system based on data mining and analysis provided by an embodiment of the present invention first obtains operation information and its corresponding initial user demand information, the operation information including sales product information, sales time, sales channel information, operation cost, operation profit, and project goal, then randomly obtains N first users corresponding to the initial user demand information from a user collection library, and randomly extracts M second users from the user collection library; sends operation information to the N first users and the M second users to obtain the response data of the N first users and the M second users to the operation information respectively, then determines the first user demand information and the second user demand information interested in the operation information according to the response data of the N first users and the M second users to the operation information respectively, and determines the final user demand information according to the initial user demand information, the first user demand information and the second user demand information; finally, sends operation information to the corresponding user according to the final user demand information; the present invention can better understand the operation status of each product and business line through the final user demand information, provide support for the optimization and improvement of decision-making, and realizes sending operation information to users interested in operation information through the present invention, thereby improving the accuracy of pushing relevant information to consumers. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 is a flow chart of an intelligent operation method based on data mining and analysis provided by an embodiment of the present invention;

[0051] Figure 2 is a flow chart of an intelligent operation method based on data mining and analysis provided by an embodiment of the present invention;

[0052] Figure 3 It is a structural diagram of an intelligent operation system based on data mining and analysis provided by an embodiment of the present invention. DETAILED DESCRIPTION

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

[0054] Embodiment 1:

[0055] See also Figure 1, is an intelligent operation method based on data mining and analysis provided by an embodiment of the present invention, the method specifically includes S101-S106:

[0056] S101, obtaining operation information and its corresponding initial user demand information, the operation information including sales product information, sales time, sales channel information, operation cost, operation profit, and project goal.

[0057] Among them, the initial user demand information is a user portrait initially drawn up based on the operational information. The user demand information may include: user gender, user age group, region, occupation, educational background, consumption habits, shopping frequency, brand preference, social media activity, etc., which is not limited in this embodiment.

[0058] It should be noted that the operation information can appear in the form of text messages, links, or mini-program advertisements, and the sales product information, sales time, sales channel information, operation costs, operation profits, and project goals can be set in the operation information. Among them, the sales product information refers to the products to be sold, the sales time and the time when the products are sold, and the sales channel refers to the way the products are sold, which can be official website sales, physical sales, or third-party platform sales.

[0059] Optionally, the operation information in this embodiment may also include data on basic resources, operation costs, operation profits, project goals, etc. By integrating and analyzing the data on basic resources, operation costs, operation profits, project goals, etc. and displaying them visually, it is possible to better understand the operation status of each product and business line, and provide support for the optimization and improvement of decision-making.

[0060] S102: randomly obtain N first users corresponding to the initial user demand information from a user collection database, and randomly select M second users from the user collection database.

[0061] Among them, the user collection library stores user demand information corresponding to a large number of users. This embodiment obtains N first users corresponding to the initial user demand information based on the user collection library, and randomly extracts M second users from the user collection library. For example, the initial user demand information requires females with an age range of 20-30, then N female users with an age range of 20-30 are randomly obtained from the user collection library, and M second users are randomly extracted from the user collection library. The second user is a user with no restrictions on user portraits, and is randomly extracted from the user collection library. In this embodiment, the values ​​of N and M can be set according to actual needs, such as N is 100 and M is 500.

[0062] In an optional embodiment provided in the present application, the randomly obtaining N first users corresponding to the initial user demand information from the user collection library, and randomly extracting M second users from the user collection library, include: determining a target user corresponding to the initial user demand information in the user collection library, and randomly extracting N first users from the target users; and randomly extracting M second users other than the target users from the user collection library, where the number M is greater than N.

[0063] S103: Send operation information to the N first users and the M second users to obtain response data of the N first users and the M second users to the operation information respectively.

[0064] The response data includes at least one or more of the following: page click behavior, page browsing time, reservation purchase behavior, interested click behavior, uninterested click behavior, etc. This embodiment does not make specific limitations on this.

[0065] S104: Determine first user demand information and second user demand information interested in the operation information according to response data of the N first users and the M second users to the operation information respectively.

[0066] In this embodiment, the demand information of the first user and the demand information of the second user who are interested in the operation information can be determined through the specific data content of the response data. Specifically, the response data of the first user and the second user to the operation information are obtained. If the user triggers the behavior of clicking on the interested behavior, the behavior of making an appointment to buy, or the browsing time exceeds the preset time in the response data, it can be determined that the user is interested in the operation information, that is, the user who may make a purchase in the future. Then, based on the obtained interested first user and second user, the demand information of the first user and the second user is determined.

[0067] In an optional embodiment provided in the present application, determining the first user demand information and the second user demand information interested in the operation information according to the response data of the N first users and the M second users to the operation information respectively includes:

[0068] S1041, performing clustering calculation on the response data of the N first users to the operation information to obtain a first cluster; performing clustering calculation on the response data of the M second users to the operation information to obtain a second cluster.

[0069] Specifically, in this embodiment, the first cluster and the second cluster can be calculated by the K-means clustering algorithm. The K-means clustering algorithm is a distance-based clustering algorithm that divides data points into K clusters through iterative calculation so that the sum of the distances from each data point to the center of its cluster is minimized.

[0070] The K-means clustering algorithm is based on the distance l and divides the response data set into Divided into k clusters, the clusters after division are , under this partition, the sum of squared error (SSE) is minimized:

[0071]

[0072] in, Represents the cluster center of each cluster, and s is the initial average round coefficient.

[0073] S1042: Determine first user demand information interested in the operation information according to the cluster center of the first cluster; and determine second user demand information interested in the operation information according to the cluster center of the second cluster.

[0074] In this embodiment, if multiple first clusters or multiple second clusters are obtained, the first user demand information and the second user demand information that the operation information is interested in can be determined according to each cluster center in the first cluster or the second cluster.

[0075] For example, there are two first clusters determined by clustering calculation, one of which has an age of 15 and the other has an age of 20, and the age in the first user demand information is 15-20 years old. That is, the user demand information is obtained by finding the union of the cluster centers in multiple clusters.

[0076] S105, determining final user demand information according to the initial user demand information, the first user demand information and the second user demand information.

[0077] It should be noted that the initial user demand information in this embodiment is the initial user demand information that is set for the operation information manually or based on experience, and the scope of the user portrait description in the initial user demand information is larger than the scope of the first user demand information. For example, the age group in the initial user demand information is set to 10-30 years old, but after N first users who meet this age group in this embodiment send operation information, it is found that only users aged 20-25 are interested in the sent operation information, that is, the initial user demand information can be further accurately positioned through this application, thereby obtaining the first user demand information.

[0078] Then, since the initial user demand information is the initial user demand information set for the operation information manually or based on experience, it may have the problem of incomplete user positioning. Therefore, this embodiment needs to determine the second user who is interested in the sent operation information based on the response data of the M second users to the operation information, and then determine the second user demand information based on the second user. Since the second user is any M users in the user collection library except the first user, the second user demand information determined by the second user can expand the user demand information for the operation information.

[0079] For example, Company A developed a blind box toy, and its initial user demand information targeted the age group of 10-25 years old. However, through the response data of M second users to the operational information, it can be determined that the age group of 25-30 years old is also interested in blind box toys, thereby expanding the user demand information age group to 10-30 years old.

[0080] In an optional embodiment provided in the present application, the final user demand information is determined based on the initial user demand information, the first user demand information and the second user demand information, including: performing clustering calculations on each field in the initial user demand information, the first user demand information and the second user demand information to obtain field clustering results; determining the corresponding user field portrait based on the cluster center in the field clustering results of each field; and combining the user field portraits of all fields into the final user demand information.

[0081] For example, if the user demand information includes age fields and income fields, clustering calculations can be performed on the age fields and income fields respectively to obtain field clustering results, and then the corresponding age and income field portraits can be determined based on the cluster centers in the field clustering results of each field; and the field portraits of all fields can be combined into the final user demand information.

[0082] In another optional embodiment provided in the present application, the final user demand information is determined based on the initial user demand information, the first user demand information and the second user demand information, including: performing clustering calculation on the first user demand information and the second user demand information, and determining the cluster center in the clustering result obtained by the clustering calculation as the intermediate user demand information; and obtaining the final user demand information by intersecting each field of the intermediate user demand information and the initial user demand information.

[0083] S106, sending the operation information to the user corresponding to the final user demand information, so that the terminal device of the user displays the operation information.

[0084] After determining the final user demand information, the corresponding target users are determined based on the final user demand information, and then the operation information is sent to the target users, thereby accurately positioning the target users, solving the problem that traditional marketing operation methods are difficult to help businesses quickly and accurately find customers and occupy the market.

[0085] An intelligent operation method based on data mining and analysis provided by an embodiment of the present invention first obtains operation information and its corresponding initial user demand information, wherein the operation information includes sales product information, sales time, and sales channel information, and then randomly obtains N first users corresponding to the initial user demand information from a user collection library, and randomly extracts M second users from the user collection library; sends the operation information to the N first users and the M second users to obtain the response data of the N first users and the M second users to the operation information respectively, and then determines the first user demand information and the second user demand information that are interested in the operation information according to the response data of the N first users and the M second users to the operation information respectively, and determines the final user demand information according to the initial user demand information, the first user demand information, and the second user demand information; finally, sends the operation information to the corresponding user according to the final user demand information. The present invention thus realizes sending operation information to users who are interested in the operation information, thereby improving the accuracy of pushing relevant information to consumers and solving the problem that traditional marketing operation methods are difficult to help businesses quickly and accurately find customers and occupy the market; and by integrating multi-dimensional data sources, using machine learning algorithms for deep mining and predictive analysis, real-time monitoring of key operation indicators, and optimizing resource allocation, decision-making efficiency is improved.

[0086] Embodiment 2:

[0087] See also Figure 2 , is an intelligent operation method based on data mining and analysis provided by an embodiment of the present invention, the method specifically includes:

[0088] S201, obtaining operation information and its corresponding initial user demand information, the operation information including sales product information, sales time, sales channel information, operation cost, operation profit, and project goal.

[0089] S202: randomly obtain N first users corresponding to the initial user demand information from the user collection library, and randomly select M second users from the user collection library.

[0090] S203: Send operation information to the N first users and the M second users to obtain response data of the N first users and the M second users to the operation information respectively.

[0091] The response data includes at least one or more of the following: page click behavior, page browsing time, reservation purchase behavior, interested click behavior, and uninterested click behavior.

[0092] S204, if the number of first users and second users interested in the operation information is greater than or equal to the corresponding preset value; determine the first user demand information and the second user demand information interested in the operation information according to the response data of N first users and M second users to the operation information respectively.

[0093] S205: Determine final user demand information according to the initial user demand information, the first user demand information and the second user demand information.

[0094] S206: Sending operation information to the corresponding user according to the final user demand information, so that the user's terminal device displays the operation information.

[0095] It should be noted that the description of steps S201 to S206 in this embodiment is the same as the description of the corresponding steps in Embodiment 1, and will not be repeated in this embodiment.

[0096] S207, if the number of first users and second users who are interested in the operation information is less than the corresponding preset value; then update the operation information according to the demand information of the first users and the second users who are interested in the operation information, and jump to the step of sending the operation information to N first users and M second users to continue executing until the number of the demand information of the first users and the second users who are interested in the operation information is greater than or equal to the corresponding preset value.

[0097] In this embodiment, if the number of the first user and the second user who are interested in the operation information is less than the corresponding preset value, it means that few users are interested in the operation information. Therefore, it is necessary to update the operation information, such as updating the sales products (such as adjusting the appearance and price of the products), sales channels, or sales time, so as to achieve the update of the operation information to better meet the needs of users or increase the interest of users in the operation information.

[0098] In an optional embodiment provided in the present application, updating the operation information according to the first user demand information and the second user demand information of interest in the operation information includes:

[0099] S2071, obtaining interest preference information corresponding to the first user demand information and the second user demand information that are interested in the operation information.

[0100] S2072, converting the interest preference information into a feature vector matrix.

[0101] Specifically, the converting of the interest preference information into a feature vector matrix includes: determining the weight value of each interest preference information based on the number of times the same preference information appears in the interest preference information; determining the interest preference information whose weight value is greater than the target weight value as the target interest preference information; and converting the target interest preference information and its corresponding weight value into the feature vector matrix.

[0102] S2073, input the feature vector matrix into the operation information prediction model to obtain corresponding operation prediction information.

[0103] The operation forecast information at least includes the forecasted sales product information, sales time, and sales channel information.

[0104] In this embodiment, the training process of the operation information prediction model is: obtaining the interest preference information of sample users and their corresponding operation annotation information, wherein the operation annotation information at least includes predicted sales product information, sales time, and sales channel information; converting the interest preference information of the sample users into a sample feature vector matrix; inputting the sample feature vector matrix into the operation information prediction model to obtain corresponding operation prediction information; calculating the loss value based on the operation prediction information and the operation annotation information; and stopping the training of the operation information prediction model when the loss value is less than a preset loss value.

[0105] S2074, update the operation information according to the operation forecast information.

[0106] Another intelligent operation method based on data mining and analysis provided by an embodiment of the present invention comprises the following steps: randomly obtaining N first users corresponding to initial user demand information from a user collection library, and randomly extracting M second users from the user collection library; sending operation information to the N first users and the M second users to obtain the response data of the N first users and the M second users to the operation information respectively; if the number of the first users and the second users interested in the operation information is less than the corresponding preset value, then updating the operation information according to the first user demand information and the second user demand information interested in the operation information; if the number of the first users and the second users interested in the operation information is greater than or equal to the corresponding preset value, then determining the first user demand information and the second user demand information interested in the operation information according to the response data of the N first users and the M second users respectively to the operation information, and determining the final user demand information according to the initial user demand information, the first user demand information and the second user demand information; finally sending the operation information to the corresponding user according to the final user demand information. Thus, the present invention realizes sending operation information to users interested in the operation information, thereby improving the accuracy of pushing relevant information to consumers, and solving the problem that traditional marketing operation methods are difficult to help merchants quickly and accurately find customers and occupy the market. By integrating multi-dimensional data sources and using machine learning algorithms to conduct in-depth mining and predictive analysis, key operational indicators can be monitored in real time and resource allocation can be optimized, thereby improving decision-making efficiency.

[0107] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0108] Embodiment 3:

[0109] In one embodiment, an intelligent operation system based on data mining and analysis is provided. Figure 3 As shown, the functional modules of the intelligent operation system based on data mining and analysis are described in detail as follows:

[0110] An acquisition module 31 is used to acquire operation information and its corresponding initial user demand information, wherein the operation information includes sales product information, sales time, sales channel information, operation cost, operation profit, and project goal;

[0111] The acquisition module 31 is further used to randomly acquire N first users corresponding to the initial user demand information from the user collection library, and randomly extract M second users from the user collection library;

[0112] The sending module 32 is used to send the operation information to the N first users and the M second users to obtain response data of the N first users and the M second users to the operation information respectively, wherein the response data includes at least one or more of the following: page click behavior, page browsing time, reservation purchase behavior, interested click behavior, and uninterested click behavior;

[0113] A determination module 33, configured to determine first user demand information and second user demand information interested in the operation information according to the response data of the N first users and the M second users to the operation information respectively;

[0114] The determination module 33 is further configured to determine final user demand information based on the initial user demand information, the first user demand information and the second user demand information;

[0115] The sending module 32 is further configured to send the operation information to the user corresponding to the final user demand information, so that the user's terminal device displays the operation information.

[0116] In an optional embodiment, the acquisition module 31 is specifically configured to:

[0117] Determine target users corresponding to the initial user demand information in the user set database, and randomly select N first users from the target users;

[0118] M second users are randomly selected from the user set library except the target user, where the number of M is greater than N.

[0119] In an optional embodiment, the determination module 33 is specifically configured to:

[0120] Performing clustering calculation on the response data of the N first users to the operation information to obtain a first cluster; performing clustering calculation on the response data of the M second users to the operation information to obtain a second cluster;

[0121] Determine first user demand information interested in the operation information according to the cluster center of the first cluster; and determine second user demand information interested in the operation information according to the cluster center of the second cluster.

[0122] In an optional embodiment, the determination module 33 is specifically configured to:

[0123] Performing clustering calculations on each field in the initial user demand information, the first user demand information, and the second user demand information to obtain a field clustering result;

[0124] The corresponding user field portrait is determined according to the cluster center in the field clustering result of each field; and the user field portraits of all fields are combined into the final user demand information.

[0125] In an optional embodiment, the determination module 33 is specifically configured to:

[0126] Performing clustering calculation on the first user demand information and the second user demand information, and determining a cluster center in a clustering result obtained by the clustering calculation as the intermediate user demand information;

[0127] The final user demand information is obtained by intersecting the fields of the intermediate user demand information and the initial user demand information.

[0128] In an optional embodiment, the intelligent operation system further includes an updating module 34, which is specifically used to:

[0129] If the number of first users and second users who are interested in the operation information is less than the corresponding preset value, the operation information is updated according to the first user demand information and the second user demand information who are interested in the operation information, and the step of sending the operation information to the N first users and the M second users is jumped to continue the execution until the number of first user demand information and second user demand information who are interested in the operation information is greater than or equal to the corresponding preset value.

[0130] In an optional embodiment, the updating module 34 is specifically configured to:

[0131] Obtaining interest preference information corresponding to the first user demand information and the second user demand information of interest to the operation information;

[0132] Converting the interest preference information into a feature vector matrix;

[0133] Inputting the characteristic vector matrix into the operation information prediction model to obtain corresponding operation prediction information, wherein the operation prediction information at least includes predicted sales product information, sales time, and sales channel information;

[0134] The operation information is updated according to the operation forecast information.

[0135] In an optional embodiment, the updating module 34 is specifically configured to:

[0136] Determining a weight value of each interest preference information based on the number of times the same preference information appears in the interest preference information;

[0137] Determine the interest preference information whose weight value is greater than the target weight value as the target interest preference information;

[0138] The target interest preference information and its corresponding weight value are converted into the feature vector matrix.

[0139] In an optional embodiment, the intelligent operation system further includes a training module 35 for training the operation information prediction model, which is specifically used for:

[0140] Obtaining sample user interest preference information and corresponding operation annotation information, wherein the operation annotation information at least includes predicted sales product information, sales time, and sales channel information;

[0141] Converting the interest preference information of the sample users into a sample feature vector matrix;

[0142] Inputting the sample feature vector matrix into the operation information prediction model to obtain corresponding operation prediction information;

[0143] Calculating a loss value according to the operation prediction information and the operation annotation information;

[0144] When the loss value is less than a preset loss value, the training of the operation information prediction model is stopped.

[0145] It should be noted that the above detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the art to which the present application belongs.

Claims

1. An intelligent operation method based on data mining and analysis, characterized in that: The method comprises: Obtaining operation information and its corresponding initial user demand information, wherein the operation information includes sales product information, sales time, sales channel information, operation cost, operation profit, and project goals; Randomly obtain N first users corresponding to the initial user demand information from the user collection database, and randomly select M second users from the user collection database; Sending the operation information to the N first users and the M second users to obtain response data of the N first users and the M second users to the operation information, respectively, the response data including: page click behavior, page browsing time, reservation purchase behavior, interested click behavior, and uninterested click behavior; Determining first user demand information and second user demand information interested in the operation information according to the response data of the N first users and the M second users to the operation information respectively; Determining final user demand information according to the initial user demand information, the first user demand information and the second user demand information; Sending the operation information to the corresponding user according to the final user demand information so that the user's terminal device displays the operation information; The randomly acquiring N first users corresponding to the initial user demand information from the user collection library, and randomly extracting M second users from the user collection library, includes: Determine target users corresponding to the initial user demand information in the user set database, and randomly select N first users from the target users; Randomly select M second users from the user set library except the target user, where the number of M is greater than N; The determining, according to the response data of the N first users and the M second users to the operation information respectively, the first user demand information and the second user demand information interested in the operation information comprises: Performing clustering calculation on the response data of the N first users to the operation information to obtain a first cluster; performing clustering calculation on the response data of the M second users to the operation information to obtain a second cluster; Determining first user demand information interested in the operation information according to the cluster center of the first cluster; determining second user demand information interested in the operation information according to the cluster center of the second cluster; The first cluster and the second cluster are obtained by calculating the K-means clustering algorithm. The K-means clustering algorithm is a distance-based clustering algorithm that divides data points into K clusters through iterative calculations so that the sum of the distances from each data point to the center of its cluster is minimized. The K-means clustering algorithm is based on the distance l and clusters the response data set D = {d1, d2, ...d n } is divided into k clusters, and the clusters after division are C = {c1, c2, ...c k }, under this partition, the sum of squared errors is minimal: Among them, xl i represents the cluster center of each cluster, s is the initial average round coefficient; The determining final user demand information according to the initial user demand information, the first user demand information and the second user demand information includes: Performing clustering calculation on the first user demand information and the second user demand information, and determining a cluster center in a clustering result obtained by the clustering calculation as the intermediate user demand information; Obtaining the final user demand information by intersecting the fields of the intermediate user demand information and the initial user demand information; After determining the first user demand information and the second user demand information that are interested in the operation information, the method further includes: If the number of the first user and the second user who are interested in the operation information is less than the corresponding preset value, the operation information is updated according to the first user demand information and the second user demand information who are interested in the operation information, and the step of sending the operation information to the N first users and the M second users is continued until the number of the first user demand information and the second user demand information who are interested in the operation information is greater than or equal to the corresponding preset value; The updating of the operation information according to the first user demand information and the second user demand information of interest in the operation information comprises: Obtaining interest preference information corresponding to the first user demand information and the second user demand information of interest to the operation information; Converting the interest preference information into a feature vector matrix; Inputting the characteristic vector matrix into the operation information prediction model to obtain corresponding operation prediction information, wherein the operation prediction information includes predicted sales product information, sales time, sales channel information, operation cost, operation profit, and project goal; updating the operation information according to the operation forecast information; The converting the interest preference information into a feature vector matrix includes: Determining a weight value of each interest preference information based on the number of times the same preference information appears in the interest preference information; Determine the interest preference information whose weight value is greater than the target weight value as the target interest preference information; Converting the target interest preference information and its corresponding weight value into the feature vector matrix; The training process of the operation information prediction model is as follows: Obtaining sample user interest preference information and corresponding operational annotation information, wherein the operational annotation information includes predicted sales product information, sales time, and sales channel information; Converting the interest preference information of the sample users into a sample feature vector matrix; Inputting the sample feature vector matrix into the operation information prediction model to obtain corresponding operation prediction information; Calculating a loss value according to the operation prediction information and the operation annotation information; When the loss value is less than a preset loss value, the training of the operation information prediction model is stopped.

2. The method according to claim 1, characterized in that The determining final user demand information according to the initial user demand information, the first user demand information and the second user demand information includes: Performing clustering calculations on each field in the initial user demand information, the first user demand information, and the second user demand information to obtain a field clustering result; The corresponding user field portrait is determined according to the cluster center in the field clustering result of each field; and the user field portraits of all fields are combined into the final user demand information.

3. An intelligent operation system based on data mining and analysis, the system implements the method according to any one of claims 1-2, characterized in that: The system comprises: An acquisition module is used to acquire operation information and its corresponding initial user demand information, wherein the operation information includes sales product information, sales time, sales channel information, operation cost, operation profit, and project goal; The acquisition module is further used to randomly acquire N first users corresponding to the initial user demand information from the user collection library, and randomly select M second users from the user collection library; A sending module, used for sending the operation information to the N first users and the M second users, so as to obtain response data of the N first users and the M second users to the operation information respectively, wherein the response data includes: page click behavior, page browsing time, reservation purchase behavior, interested click behavior, and uninterested click behavior; A determination module, configured to determine first user demand information and second user demand information interested in the operation information according to response data of the N first users and the M second users to the operation information respectively; The determining module is further configured to determine final user demand information based on the initial user demand information, the first user demand information and the second user demand information; The sending module is further used to send the operation information to the corresponding user according to the final user demand information, so that the user's terminal device displays the operation information.

Images