A method and system for extracting advertisement information based on automated interaction

Abstract

The present application relates to the technical field of information extraction, and in particular to a method and system for extracting advertisement delivery information based on automated interaction. The present application first obtains the interactive content between the user and the automated interactive system; then extracts the scheme identifier based on the interactive content, matches it in the stored database, and determines the delivery scheme corresponding to the scheme identifier; then extracts key information and performs text recognition on the interactive content to obtain user information; finally, structured user portrait data is generated based on the user information, and the user portrait data is classified into the corresponding delivery scheme according to the delivery scheme, and a delivery information set organized by the delivery scheme is obtained, which improves the efficiency and accuracy of chat record information extraction in investment promotion advertising delivery, and reduces labor costs and subjective errors; at the same time, the generated structured user portrait data and the information set organized by the delivery scheme provide data support for subsequent precise delivery and effect optimization.

Description

Technical Field

[0001] The present application relates to the technical field of information extraction, and in particular to a method and system for extracting advertisement delivery information based on automated interaction. Background Art

[0002] With the rapid development of the Internet and the continuous upgrading of the advertising industry, digital advertising has become an important means for enterprises to conduct marketing and promotion. The accuracy and efficiency of advertising directly affect the advertiser's return on investment and the user's experience satisfaction. How to quickly and accurately extract key information from massive user interaction data, conduct user profiling and delivery optimization is a problem that needs to be solved urgently in the field of digital advertising.

[0003] The investment promotion advertising resources include some resources obtained through chats between robots and users. It is necessary to extract the customer's contact information, name, geographical location and other information from the chat content. The existing technology uses manual processing, which is inefficient and the quality of the results cannot be guaranteed. This situation needs to be further improved. Summary of the invention

[0004] In order to solve the problem of low efficiency in the existing advertisement placement information extraction, the present application provides an advertisement placement information extraction method and system based on automated interaction, which adopts the following technical solutions:

[0005] In a first aspect, the present application provides an advertisement delivery information extraction method based on automated interaction, comprising the following steps:

[0006] Obtain the interaction content between the user and the automated interactive system;

[0007] Extracting a scheme identifier according to the interactive content;

[0008] According to the scheme identifier, matching is performed in a stored database to determine the delivery scheme corresponding to the scheme identifier;

[0009] Extract key information and perform text recognition on the interactive content to obtain user information;

[0010] According to the user information, structured user portrait data is generated, and according to the delivery plan, the user portrait data is classified into a corresponding delivery plan to obtain a delivery information set organized according to the delivery plan.

[0011] By adopting the above technical scheme, the present application first obtains the interaction content between the user and the automated interaction system; then, extracts the scheme identifier according to the interaction content; then matches the stored database according to the scheme identifier to determine the delivery scheme corresponding to the scheme identifier; then performs key information extraction and text recognition on the interaction content to obtain user information; finally, based on the user information, structured user portrait data is generated, and according to the delivery scheme, the user portrait data is classified into the corresponding delivery scheme to obtain a delivery information set organized by the delivery scheme, and through automated information extraction and matching, the efficiency and accuracy of chat record information extraction in investment promotion advertising delivery are improved, and labor costs and subjective errors are reduced; at the same time, the generated structured user portrait data and the information set organized by the delivery scheme provide high-quality data support for subsequent precise delivery and effect optimization, and improve the overall efficiency and level of delivery management.

[0012] Optionally, extracting the scheme identifier according to the interactive content specifically includes the following steps:

[0013] Obtaining a page link in the interactive content;

[0014] Parsing the page link to obtain page link parameters;

[0015] According to preset extraction rules, the scheme identifier is extracted from the page link parameters, and the extraction rules include: judging in sequence according to a preset priority order whether there are parameters of a preset type in the page link parameters, the parameters of the preset type include numeric parameters and string parameters, and the numeric parameters and the string parameters can uniquely identify the delivery scheme.

[0016] By adopting the above technical scheme, the present application first obtains the page link in the interactive content; then parses the page link to obtain the page link parameters; then, according to the preset extraction rules, extracts the scheme identifier from the page link parameters; wherein the extraction rules include: judging in sequence according to the preset priority order whether there are parameters of preset types in the page link parameters, the parameters of preset types include numeric parameters and string parameters, the numeric parameters and string parameters can uniquely identify the delivery scheme, and extracting the scheme identifier by parsing the page link parameters avoids the difficulty of directly identifying the scheme identifier from complex and diverse interactive content, thereby improving the accuracy and efficiency of extraction. At the same time, the extraction rules judged by the preset priority order can flexibly cope with page link parameters of different formats and types, and have strong adaptability and robustness.

[0017] Optionally, performing key information extraction and text recognition on the interactive content to obtain user information specifically includes the following steps:

[0018] Identifying key questions and user answers from the interactive content, wherein the key questions include name-related questions and contact information-related questions;

[0019] The user's answer to the name-related question is analyzed according to the preset name judgment rule to obtain the user's name information; the user's answer to the contact information-related question is analyzed according to the preset contact information judgment rule to obtain the user's contact information;

[0020] The interaction time information and IP address information are extracted from the interaction content, and the region is determined based on the IP address information to obtain the user region information.

[0021] By adopting the above technical solution, the present application can effectively respond to user inputs of different expressions and formats by analyzing user answers using preset name and contact information judgment rules, thereby improving the accuracy and adaptability of information extraction; at the same time, by comprehensively utilizing metadata information such as interaction time and IP address, the user's geographical attributes can be obtained, enriching the dimensionality and comprehensiveness of user information; and finally forming structured user information.

[0022] Optionally, after obtaining the delivery information set organized according to the delivery plan, the method further comprises the following steps: obtaining an original data file generated by an automated interactive system corresponding to the delivery plan, the original data file comprising historical interaction records with multiple users;

[0023] Parsing the original data file to extract data related to delivery effect, wherein the data related to delivery effect includes landing page information, interaction time information, IP address information, IP address type information, interaction entry source information and keyword information;

[0024] According to the delivery effect related data, statistically analyzing the effect data of the delivery plan, wherein the effect data includes the number of chat conversations, whether user information is obtained, the amount of user information obtained, the conversation efficiency, and the statistical date;

[0025] An effect report of the delivery plan is generated based on the effect data, and the effect report is provided to the advertising delivery personnel for delivery effect analysis.

[0026] By adopting the above technical solutions, this application directly obtains and parses the original data files of the automated interactive system to comprehensively and accurately extract data related to the delivery effect, thus avoiding the limitations and errors of data acquisition; the data related to the delivery effect covers multiple dimensions such as landing page, interaction time, IP address, source, keywords, etc., providing rich data support for the comprehensive evaluation of the delivery effect; the effect data generated based on the statistics of effect-related data, including key indicators such as the number of chat conversations, user information acquisition, and conversation efficiency, intuitively quantifies the actual effect of the delivery plan, and the automatically generated effect report provides advertisers with an intuitive and easy-to-use analysis tool, which is convenient for advertisers to find problems, optimize strategies, and improve delivery ROI.

[0027] Optionally, after collecting statistics on the effect data of the delivery plan according to the delivery effect-related data, the method further includes the following steps:

[0028] Performing cluster analysis on the data related to the delivery effect, and dividing the users into different user groups according to user characteristic parameters, wherein the user characteristic parameters include geographical distribution, interaction time distribution, and keyword preference;

[0029] For different user groups, the effect data of the corresponding delivery plans are counted respectively to obtain the delivery effect data of each user group; the delivery effect data of each user group are weighted summed to obtain the comprehensive effect score of each user group;

[0030] Calculate the ratio of the comprehensive effect score of each user group to the total comprehensive effect scores of all user groups, and use the ratio as the contribution of the corresponding user group to the delivery effect;

[0031] Sorting each user group according to the contribution of the delivery effect, and identifying the key user group according to the sorting result;

[0032] According to the distribution law of the user characteristic parameters of the key user group, the characteristic portrait of the key user group is summarized and added to the effect report of the delivery plan.

[0033] By adopting the above technical solution, in order to further explore the differences in the contributions of different user groups to the delivery effect, this application conducts cluster analysis and group division on users to discover the differences and key influencing factors in the delivery effect of different user groups; comprehensively considers multi-dimensional user characteristic parameters such as region, time, keywords, etc., and comprehensively characterizes the attributes and preferences of the user group, and then introduces the calculation of comprehensive effect scores and contribution degrees to quantitatively evaluate the contribution of different user groups to the overall delivery effect, which helps to reasonably allocate delivery resources and optimize strategies; finally, automatically identifies key user groups and summarizes feature portraits, which can help advertisers quickly lock in high-value users and optimize target group selection and creative design.

[0034] Optionally, before sorting each user group according to the delivery effect contribution, the method further includes the following steps:

[0035] Perform time weighting on the delivery effect contribution of each user group to obtain the adjusted delivery effect contribution;

[0036] The contribution of the adjusted delivery effect Among them, c it It represents the contribution of the i-th user group to the delivery effect in the t-th time period. T represents the total time span of the delivery effect data, that is, the number of time periods from the start of delivery to the current time point. α represents the time attenuation factor, 0<α≤1.

[0037] By adopting the above technical solution, in order to more accurately evaluate the delivery effect contribution of the user group, this application takes into account the time decay characteristics of the user group contribution, makes weighted adjustments to the contributions of different time periods, and gives higher weights to recent data to obtain the adjusted delivery effect contribution, thereby better reflecting the user group's immediate contribution and weight changes, thereby quickly discovering user groups with outstanding recent performance and significantly improved contribution, and timely optimizing delivery strategies and resource allocation.

[0038] In a second aspect, the present application provides an advertisement delivery information extraction system based on automated interaction, comprising:

[0039] An interactive content acquisition module is used to acquire the interactive content between the user and the automated interactive system;

[0040] A scheme identifier extraction module, used for extracting a scheme identifier according to the interactive content;

[0041] A delivery plan determination module, used to match the delivery plan corresponding to the plan identifier in the stored database according to the plan identifier;

[0042] A user information acquisition module, used to extract key information and recognize text from the interactive content to obtain user information;

[0043] The delivery information set acquisition module is used to generate structured user portrait data based on the user information, and classify the user portrait data into corresponding delivery plans according to the delivery plans to obtain a delivery information set organized according to the delivery plans.

[0044] Optionally, the scheme identification extraction module includes:

[0045] A page link acquisition unit, used to acquire a page link in the interactive content;

[0046] A link parameter acquisition unit, used for parsing the page link to obtain page link parameters;

[0047] A scheme identifier extraction unit is used to extract the scheme identifier from the page link parameter according to a preset extraction rule, wherein the extraction rule includes:

[0048] It is determined in sequence according to a preset priority order whether there are parameters of a preset type in the page link parameters, the parameters of the preset type include numeric parameters and string parameters, and the numeric parameters and the string parameters can uniquely identify the delivery plan.

[0049] Optionally, the user information acquisition module includes:

[0050] An information identification unit, used to identify key questions and user answers from the interactive content, wherein the key questions include name-related questions and contact information-related questions;

[0051] A name information acquisition unit, used to analyze the user's answer to the name-related question according to a preset name judgment rule to obtain the user's name information;

[0052] A contact information acquisition unit, configured to analyze the user's answer to the contact information-related question according to a preset contact information judgment rule to obtain the user's contact information;

[0053] The region information acquisition unit is used to extract the interaction time information and IP address information from the interaction content, and perform region judgment according to the IP address information to obtain user region information.

[0054] Optionally, the system further includes:

[0055] An original data file acquisition module, used to acquire an original data file generated by an automated interactive system corresponding to the delivery plan, wherein the original data file includes historical interaction records with multiple users;

[0056] A delivery effect related data extraction module is used to parse the original data file and extract delivery effect related data, wherein the delivery effect related data includes landing page information, interaction time information, IP address information, IP address type information, interaction entry source information and keyword information;

[0057] An effect data statistics module, for counting effect data of the delivery scheme according to the delivery effect related data, wherein the effect data includes the number of chat conversations, whether user information is obtained, the amount of user information obtained, conversation efficiency and statistical date;

[0058] The effect report generation module is used to generate an effect report of the delivery plan according to the effect data, and provide the effect report to the advertising delivery personnel for delivery effect analysis.

[0059] Optionally, the system further includes:

[0060] A user group division module, used to perform cluster analysis on the delivery effect related data and divide users into different user groups according to user characteristic parameters, wherein the user characteristic parameters include geographical distribution, interaction time distribution and keyword preference;

[0061] The user group delivery effect data acquisition module is used to collect statistics on the effect data of the corresponding delivery plans for different user groups and obtain the delivery effect data of each user group;

[0062] The user group effect score calculation module is used to perform weighted summation on the delivery effect data of each user group to obtain the comprehensive effect score of each user group;

[0063] The delivery effect contribution determination module is used to calculate the ratio of the comprehensive effect score of each user group to the total comprehensive effect scores of all user groups, and use the ratio as the delivery effect contribution of the corresponding user group;

[0064] A key user group identification module is used to sort each user group according to the contribution of the delivery effect, and identify the key user group according to the sorting result;

[0065] The key user group feature portrait adding module is used to summarize the feature portrait of the key user group according to the distribution law of the user feature parameters of the key user group, and add it to the effect report of the delivery plan.

[0066] Optionally, the system further includes:

[0067] The delivery effect contribution adjustment module is used to perform time weighting on the delivery effect contribution of each user group to obtain the adjusted delivery effect contribution;

[0068] The adjusted contribution of the delivery effect Among them, c it It represents the contribution of the i-th user group to the delivery effect in the t-th time period. T represents the total time span of the delivery effect data, that is, the number of time periods from the start of delivery to the current time point. α represents the time attenuation factor, 0<α≤1.

[0069] In summary, the present application includes at least one of the following beneficial technical effects:

[0070] 1. This application first obtains the interactive content between the user and the automated interactive system; then, extracts the solution identifier based on the interactive content; then, matches the solution identifier in the stored database to determine the delivery solution corresponding to the solution identifier; then, extracts key information and recognizes text from the interactive content to obtain user information; finally, generates structured user portrait data based on the user information, and classifies the user portrait data into the corresponding delivery solution based on the delivery solution to obtain a delivery information set organized by the delivery solution, and improves the efficiency and accuracy of chat record information extraction in investment promotion advertising delivery through automated information extraction and matching, thereby reducing labor costs and subjective errors; at the same time, the generated structured user portrait data and the information set organized by the delivery solution provide high-quality data support for subsequent precise delivery and effect optimization, and improve the overall efficiency and level of delivery management;

[0071] 2. This application directly obtains and analyzes the original data files of the automated interactive system to comprehensively and accurately extract data related to the delivery effect, thus avoiding the limitations and errors of data acquisition; the delivery effect related data covers multiple dimensions such as landing pages, interaction time, IP addresses, sources, keywords, etc., providing rich data support for the comprehensive evaluation of delivery effects; the effect data generated based on the statistics of effect-related data, including key indicators such as the number of chat conversations, user information acquisition, and conversation efficiency, intuitively quantifies the actual effect of the delivery plan, and the automatically generated effect report provides an intuitive and easy-to-use analysis tool for advertising delivery personnel, which is convenient for advertising delivery personnel to find problems, optimize strategies, and improve delivery ROI;

[0072] 3. This application discovers the differences and key influencing factors of the delivery effects of different user groups through cluster analysis and group division of users; comprehensively considers multi-dimensional user characteristic parameters such as region, time, keywords, etc., comprehensively describes the attributes and preferences of the user group, and then introduces the calculation of comprehensive effect scores and contribution degrees to quantitatively evaluate the contribution of different user groups to the overall delivery effect, thereby helping to reasonably allocate delivery resources and optimize strategies; finally, automatically identifies key user groups and summarizes feature portraits, which can help advertisers quickly lock in high-value users and optimize target group selection and creative design. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 It is a flowchart of an advertisement delivery information extraction method based on automated interaction according to an embodiment of the present application;

[0074] Figure 2 It is a flowchart of step S120 in a method for extracting advertisement delivery information based on automated interaction according to an embodiment of the present application;

[0075] Figure 3It is a flowchart of step S140 in a method for extracting advertisement delivery information based on automated interaction according to an embodiment of the present application;

[0076] Figure 4 It is a flowchart of step S150 in an advertisement delivery information extraction method based on automated interaction according to an embodiment of the present application;

[0077] Figure 5 It is a flow chart of dividing user groups in a method for extracting advertisement delivery information based on automated interaction according to an embodiment of the present application;

[0078] Figure 6 It is a module schematic diagram of an advertisement delivery information extraction system based on automated interaction according to an embodiment of the present application;

[0079] Figure 7 It is a diagram of the internal structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0080] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification and appended claims of the present application, the singular expressions "one", "a kind of", "said", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more listed items.

[0081] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0082] The embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0083] In a first aspect, the present application provides an advertisement delivery information extraction method based on automated interaction, referring to Figure 1 , including the following steps:

[0084] S110: Obtain interaction content between the user and the automated interaction system.

[0085] In this embodiment, the automated interactive system refers to a software system based on artificial intelligence technology that can conduct natural language dialogues with users and complete specific tasks, such as chat robots, intelligent customer service, etc. The interactive content refers to the dialogue records between users and the system, including the user's questions or requests, and the system's answers or feedback.

[0086] Specifically, a data collection module is set up in the automated interactive system to record and store each user session in real time; historical session records are regularly exported in batches from the system's log files or databases, where the acquired interactive content is usually in text form and carries metadata such as session ID and timestamp.

[0087] S120: Extract a solution identifier according to the interactive content.

[0088] In this embodiment, the scheme identifier refers to an identification code or number that can uniquely identify an advertising delivery scheme, and is used to associate the interactive content with a specific delivery scheme. The scheme identifier is usually embedded in the information sent by the automated interactive system, such as a welcome message, a promotion link, etc.

[0089] Specifically, firstly, a segment containing a scheme identifier is located in the interactive content by means of keyword matching, regular expressions, etc.; then, a complete scheme identifier string is extracted from the matched segment according to a predefined rule or pattern.

[0090] S130: According to the scheme identifier, a match is performed in the stored database to determine the delivery scheme corresponding to the scheme identifier.

[0091] In this embodiment, the stored database refers to a structured data set used to manage and store advertising delivery plan information, wherein each record corresponds to a delivery plan and includes metadata such as plan identification, name, delivery channel, and target population.

[0092] Specifically, the scheme identifier matching can be implemented using the index or query function of the database. For example, a SQL statement is executed to query the details of the delivery scheme corresponding to the identifier. The matching process can add necessary exception handling, such as handling the situation where the identifier does not exist.

[0093] S140: Extract key information and perform text recognition on the interactive content to obtain user information.

[0094] In this embodiment, key information refers to information points that play an important role in judging and classifying user attributes, such as name, contact information, intentions and needs, etc.; text recognition refers to accurately extracting and converting structured key information from the user's natural language expression; user information refers to a collection of a series of extracted key information, which reflects the basic characteristics of the user.

[0095] Specifically, named entity recognition, keyword matching and other technologies in natural language processing are used to identify information elements such as names, telephone numbers, email addresses, addresses, etc. in the interactive content; intention needs are obtained through text classification, keyword extraction and other methods to judge the interest preferences and purchasing intentions expressed in the user conversation.

[0096] S150. Generate structured user portrait data based on user information, and classify the user portrait data into corresponding delivery plans according to the delivery plans, so as to obtain a delivery information set organized according to the delivery plans.

[0097] In this embodiment, the user portrait refers to a comprehensive and structured characterization of a user. The delivery information set refers to a portrait set of all users reached by a delivery plan.

[0098] Specifically, based on the extracted user information, a user portrait is generated in combination with a predefined user portrait template and existing user master data, and the user portrait is classified into a delivery plan. A delivery plan identification field can be added to the user portrait data, or an association table between the user portrait and the delivery plan can be established. The classified delivery information set is stored in an independent data mart and generated in real time as a logical view.

[0099] In one embodiment, referring to Figure 2 In step S120, extracting a solution identifier according to the interactive content specifically includes the following steps:

[0100] S121. Obtain a page link in the interactive content.

[0101] In this embodiment, the page link refers to a hyperlink to an external web page contained in a message sent by the automated interaction system to the user, including an advertisement landing page, an activity promotion page, etc., which contains parameter information related to the scheme identifier.

[0102] Specifically, first use regular expressions or other pattern matching methods to search for URL-formatted strings in the text of the interactive content; then, perform format verification and normalization on the matched URLs to extract complete and valid link addresses. If an interactive content contains multiple links, filter them according to predefined rules, such as selecting the first or last link or a link at a specific location.

[0103] S122. Parse the page link to obtain page link parameters.

[0104] In this embodiment, the page link parameter refers to the part after the "?" in the URL, which consists of one or more "key=value" pairs, and different parameters are separated by "&". These parameters usually contain some information related to advertising, such as source channel, activity number, creative ID, etc.

[0105] Specifically, according to the standard format of URL, the link is divided into different parts such as protocol, domain name, path, parameters, etc.; then, the parameter part is further split and divided into multiple independent key-value pairs according to the "&" sign; finally, the parsed key-value pairs are stored as a hash table or dictionary structure to facilitate subsequent query and extraction.

[0106] S123. Extracting a solution identifier from the page link parameters according to a preset extraction rule.

[0107] In this embodiment, the extraction rules refer to some predefined mapping relationships or patterns, which are used to identify and extract the solution identifier from the page link parameters.

[0108] The extraction rules include: judging in sequence according to a preset priority order whether there are parameters of preset types in the page link parameters, the parameters of preset types include numeric parameters and string parameters, and the numeric parameters and string parameters can uniquely identify the delivery plan.

[0109] Specifically, the following method can be used to extract the solution identifier: first, try to parse a pure number of more than 8 digits from the chat page link. This pure number of more than 8 digits has data that identifies which project solution it is. We can call it the project solution identifier. If it cannot be parsed, then according to the preset parameter names such as "arg", "consult_id", etc., search for the corresponding key-value pair in the parsed link parameters; if a matching parameter name is found, its corresponding parameter value is used as the solution identifier; if no exact matching parameter name is found, you can try to use pattern matching methods such as "is", "c", etc. If these two parameters exist at the same time, take the "is" parameter as the project solution identifier.

[0110] In one embodiment, referring to Figure 3 In step S140, key information extraction and text recognition are performed on the interactive content to obtain user information, which specifically includes the following steps:

[0111] S141. Identify key questions and user answers from the interactive content, where the key questions include name-related questions and contact information-related questions.

[0112] S142: Analyze the user's answer to the name-related question according to a preset name judgment rule to obtain the user's name information.

[0113] Specifically, the name judgment rule includes judging whether it is a valid name based on the number of digits in the field answered by the user and the corresponding surname.

[0114] S143: Analyze the user's answers to the contact information-related questions according to preset contact information judgment rules to obtain the user's contact information.

[0115] Specifically, determine whether it is a mobile phone number based on the number of digits in the field answered by the user, and if so, further determine based on the format of the mobile phone number; determine whether it is a WeChat account based on the content of the field answered by the user, and if so, further determine based on the format of the WeChat account.

[0116] S144: extract interaction time information and IP address information from the interaction content, and perform region determination based on the IP address information to obtain user region information.

[0117] In one embodiment, referring to Figure 4 , step S150, after obtaining the delivery information set organized according to the delivery plan, the method further includes the following steps:

[0118] S210: Acquire an original data file generated by an automated interactive system corresponding to the delivery plan, where the original data file includes historical interaction records with multiple users.

[0119] In this embodiment, the original data file refers to a log file or data export file automatically recorded and stored by the automated interaction system during operation, which contains a large amount of user interaction history data, such as request time, request parameters, response content, etc.

[0120] S220: parse the original data file and extract data related to the delivery effect.

[0121] Among them, the data related to the delivery effect includes landing page information, interaction time information, IP address information, IP address type information, interaction entry source information and keyword information.

[0122] S230: Count the effect data of the delivery plan based on the delivery effect related data.

[0123] Among them, the effect data includes the number of chat conversations, whether user information is obtained, the amount of user information obtained, the conversation efficiency and the statistical date.

[0124] In this embodiment, the effect data refers to a series of numerical indicators obtained by summarizing and calculating the data related to the delivery effect, which is used to quantitatively evaluate the effect of advertising delivery. These indicators reflect the performance of the delivery plan in terms of audience coverage, interactive conversion, information collection, etc. from different angles, such as the number of chat conversations between users and the system, whether the user's contact information is successfully obtained, the amount of valid user information obtained, the effectiveness of the conversation, etc.

[0125] S240: Generate an effect report of the delivery plan based on the effect data, and provide the effect report to the advertising delivery personnel for delivery effect analysis.

[0126] In one embodiment, referring to Figure 5 In step S240, after calculating the effect data of the delivery plan according to the delivery effect related data, the method further includes the following steps:

[0127] S310: Perform cluster analysis on the data related to the delivery effect, and divide the users into different user groups according to user characteristic parameters.

[0128] Among them, user characteristic parameters include geographical distribution, interaction time distribution and keyword preference.

[0129] In this embodiment, cluster analysis refers to automatically classifying similar objects into the same group by calculating the similarity between data objects, thereby achieving automatic classification of users.

[0130] Specifically, based on the data related to the delivery effect, we extract the user's region, interaction time, keywords and other feature fields, select the appropriate clustering algorithm and parameters, input the user feature data into the algorithm for training and optimization, and obtain the final clustering result; finally, we evaluate the clustering results and count the number of users and feature distribution of each group.

[0131] S320: For different user groups, respectively collect statistics on the effect data of the corresponding delivery plans to obtain delivery effect data for each user group.

[0132] In this embodiment, the group effect statistics refers to further subdividing the delivery effect data based on user clustering, and calculating the independent effect indicators of each user group, such as the average number of conversations, user information collection rate, conversation efficiency, etc. of different groups. Through the group effect statistics, the contribution of different user groups to the delivery effect is compared, and high-value and low-effect groups are identified, so as to take targeted optimization measures.

[0133] Specifically, the user clustering results are associated with the data related to the delivery effect, and a group label is given to each user record; then, the data is grouped according to the group label, and the indicators of users in each group are counted and calculated to obtain the effect data such as the number of chat conversations, user information collection rate, conversation efficiency, etc. of each group; finally, the effect data of each group is summarized and compared and analyzed, and the mean of the effect indicators of each group is calculated.

[0134] S330: Perform weighted summation on the delivery effect data of each user group to obtain a comprehensive effect score of each user group.

[0135] In this embodiment, the comprehensive effect score refers to combining multiple effect indicators of each group into a comprehensive score by weighted average to reflect the overall effect performance of the group. The weight coefficient is set according to the importance of the indicator and the business goal, such as the weight of conversation efficiency is higher than the number of conversations, and the weight of user information collection rate is higher than IP address coverage.

[0136] S340, calculating the ratio of the comprehensive effect score of each user group to the total comprehensive effect scores of all user groups, and using the ratio as the delivery effect contribution of the corresponding user group.

[0137] In this embodiment, the delivery effect contribution refers to the percentage of the comprehensive effect score of a group in the total score of all groups, reflecting the contribution of the group to the overall delivery effect.

[0138] S350, sorting each user group according to the contribution of the delivery effect, and identifying the key user group according to the sorting result.

[0139] S360: Summarize the characteristic portraits of key user groups based on the distribution patterns of their user characteristic parameters and add them to the effectiveness report of the delivery plan.

[0140] In one embodiment, step S330, before weighted summing of the delivery effect data of each user group, further includes adjusting the weight of the user characteristic parameter. The adjusted user characteristic parameter weight Where n represents the number of user groups, p i represents the contribution of the delivery effect of the i-th user group, s ik It represents the similarity between the i-th user group and the k-th user group, and it represents the original weight of the j-th user feature parameter in the k-th user group.

[0141] In one embodiment, before sorting each user group according to the delivery effect contribution in step S350, the method further includes performing time weighting on the delivery effect contribution of each user group to obtain an adjusted delivery effect contribution; the adjusted delivery effect contribution Among them, cit It represents the contribution of the ith user group to the delivery effect in the tth time period, T represents the total time span of the delivery effect data, that is, the number of time periods from the start of delivery to the current time point, α represents the time decay factor, 0<α≤1. By weighted adjustment of the contribution of different time periods, a higher weight is given to recent data, and the adjusted delivery effect contribution is obtained, which better reflects the immediate contribution and weight change of the user group, so as to quickly discover the user group with outstanding recent performance and significantly improved contribution, and timely optimize the delivery strategy and resource allocation.

[0142] It should be understood that the size of the serial numbers of the steps in the above embodiments does not 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 embodiments of the present application.

[0143] In a second aspect, the present application provides an advertisement placement information extraction system based on automated interaction. The advertisement placement information extraction system based on automated interaction of the present application is described below in combination with the above-mentioned advertisement placement information extraction method based on automated interaction.

[0144] Reference Figure 6 , an advertisement delivery information extraction system based on automated interaction, comprising:

[0145] An interactive content acquisition module is used to acquire the interactive content between the user and the automated interactive system;

[0146] A scheme identifier extraction module, used for extracting a scheme identifier according to the interactive content;

[0147] A delivery plan determination module is used to match the delivery plan corresponding to the plan identifier in the stored database according to the plan identifier;

[0148] A user information acquisition module is used to extract key information and recognize text from interactive content to obtain user information;

[0149] The delivery information set acquisition module is used to generate structured user portrait data based on user information, and classify the user portrait data into corresponding delivery plans according to the delivery plans, so as to obtain a delivery information set organized according to the delivery plans.

[0150] In an optional embodiment, the scheme identification extraction module includes:

[0151] A page link acquisition unit, used to acquire page links in interactive content;

[0152] A link parameter acquisition unit, used to parse the page link and obtain the page link parameters;

[0153] The scheme identifier extraction unit is used to extract the scheme identifier from the page link parameter according to a preset extraction rule, and the extraction rule includes:

[0154] It is determined in turn according to the preset priority order whether there are parameters of preset types in the page link parameters. The parameters of preset types include numeric parameters and string parameters. The numeric parameters and string parameters can uniquely identify the delivery plan.

[0155] In an optional embodiment, the user information acquisition module includes:

[0156] An information identification unit, used to identify key questions and user answers from the interactive content, wherein the key questions include name-related questions and contact information-related questions;

[0157] A name information acquisition unit is used to analyze the user's answer to name-related questions according to preset name judgment rules to obtain the user's name information;

[0158] A contact information acquisition unit is used to analyze the user's answers to contact information related questions according to preset contact information judgment rules to obtain the user's contact information;

[0159] The region information acquisition unit is used to extract interaction time information and IP address information from the interaction content, and perform region judgment based on the IP address information to obtain user region information.

[0160] In an optional embodiment, the system further includes:

[0161] An original data file acquisition module is used to acquire an original data file generated by an automated interactive system corresponding to the delivery plan, wherein the original data file includes historical interaction records with multiple users;

[0162] The delivery effect related data extraction module is used to parse the original data file and extract delivery effect related data. The delivery effect related data includes landing page information, interaction time information, IP address information, IP address type information, interaction entry source information and keyword information;

[0163] The effect data statistics module is used to count the effect data of the delivery plan based on the relevant data of the delivery effect. The effect data includes the number of chat conversations, whether user information is obtained, the amount of user information obtained, the conversation efficiency and the statistical date; the effect report generation module is used to generate the effect report of the delivery plan based on the effect data, and provide the effect report to the advertising delivery personnel for delivery effect analysis.

[0164] In an optional embodiment, the system further includes:

[0165] The user group segmentation module is used to perform cluster analysis on data related to delivery effects and divide users into different user groups based on user characteristic parameters, including geographical distribution, interaction time distribution, and keyword preference;

[0166] The user group delivery effect data acquisition module is used to collect statistics on the effect data of the corresponding delivery plans for different user groups and obtain the delivery effect data of each user group;

[0167] The user group effect score calculation module is used to perform weighted summation on the delivery effect data of each user group to obtain the comprehensive effect score of each user group;

[0168] The delivery effect contribution determination module is used to calculate the ratio of the comprehensive effect score of each user group to the total comprehensive effect scores of all user groups, and use the ratio as the delivery effect contribution of the corresponding user group;

[0169] The key user group identification module is used to sort each user group according to the contribution of the delivery effect, and identify the key user group based on the sorting results;

[0170] The key user group feature portrait adding module is used to summarize the feature portraits of key user groups based on the distribution rules of their user feature parameters, and add them to the effect report of the delivery plan.

[0171] In an optional embodiment, the system further includes:

[0172] The delivery effect contribution adjustment module is used to perform time weighting on the delivery effect contribution of each user group to obtain the adjusted delivery effect contribution;

[0173] Adjusted delivery effect contribution Among them, c it It represents the contribution of the i-th user group to the delivery effect in the t-th time period. T represents the total time span of the delivery effect data, that is, the number of time periods from the start of delivery to the current time point. α represents the time attenuation factor, 0<α≤1.

[0174] In one embodiment, the present application provides an electronic device, which may be a server, and its internal structure diagram may be as follows: Figure 7As shown. The electronic device includes a processor, a memory and a network interface connected via a system bus. The processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the electronic device is used to store data. The network interface of the electronic device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for extracting advertising information based on automated interaction is implemented.

[0175] Those skilled in the art will understand that Figure 7 The structure shown in the figure is merely a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the electronic device to which the scheme of the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0176] In one embodiment, an electronic device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0177] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the above-mentioned computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0178] 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 method for extracting advertisement information based on automated interaction, characterized in that: The steps include: Obtain the interaction content between the user and the automated interactive system; Extracting a scheme identifier according to the interactive content; According to the scheme identifier, matching is performed in a stored database to determine the delivery scheme corresponding to the scheme identifier; Extract key information and perform text recognition on the interactive content to obtain user information; Generate structured user portrait data according to the user information, and classify the user portrait data into corresponding delivery plans according to the delivery plans to obtain a delivery information set organized according to the delivery plans; Acquire an original data file generated by an automated interactive system corresponding to the delivery plan, wherein the original data file includes historical interaction records with multiple users; Parsing the original data file to extract data related to delivery effect, wherein the data related to delivery effect includes landing page information, interaction time information, IP address information, IP address type information, interaction entry source information and keyword information; According to the delivery effect related data, statistically analyzing the effect data of the delivery plan, wherein the effect data includes the number of chat conversations, whether user information is obtained, the amount of user information obtained, the conversation efficiency, and the statistical date; Generate an effect report of the delivery plan based on the effect data, and provide the effect report to the advertising delivery personnel for delivery effect analysis; Performing cluster analysis on the data related to the delivery effect, and dividing the users into different user groups according to user characteristic parameters, wherein the user characteristic parameters include geographical distribution, interaction time distribution, and keyword preference; For different user groups, the effect data of the corresponding delivery plans are counted respectively to obtain the delivery effect data of each user group; Perform weighted summation on the delivery effect data of each user group to obtain the comprehensive effect score of each user group; Calculate the ratio of the comprehensive effect score of each user group to the total comprehensive effect scores of all user groups, and use the ratio as the delivery effect contribution of the corresponding user group; Sorting each user group according to the contribution of the delivery effect, and identifying the key user group according to the sorting result; According to the distribution law of the user characteristic parameters of the key user group, the characteristic portrait of the key user group is summarized and added to the effect report of the delivery plan.

2. The method for extracting advertisement placement information based on automated interaction according to claim 1, characterized in that: According to the interactive content extraction scheme identifier, the following steps are specifically included: Obtaining a page link in the interactive content; Parsing the page link to obtain page link parameters; The scheme identifier is extracted from the page link parameter according to a preset extraction rule, wherein the extraction rule includes: It is determined in sequence according to a preset priority order whether there are parameters of a preset type in the page link parameters, the parameters of the preset type include numeric parameters and string parameters, and the numeric parameters and the string parameters can uniquely identify the delivery plan.

3. The method for extracting advertisement placement information based on automated interaction according to claim 1, characterized in that: Extracting key information and identifying text from the interactive content to obtain user information specifically includes the following steps: Identifying key questions and user answers from the interactive content, wherein the key questions include name-related questions and contact information-related questions; Analyze the user's answer to the name-related question according to the preset name judgment rule to obtain the user's name information; The user's answers to the contact information-related questions are analyzed according to the preset contact information judgment rules to obtain the user's contact information; The interaction time information and IP address information are extracted from the interaction content, and the region is determined based on the IP address information to obtain the user region information.

4. The method for extracting advertisement placement information based on automated interaction according to claim 1, characterized in that: Before sorting each user group according to the delivery effect contribution, the method further includes the following steps: Perform time weighting on the delivery effect contribution of each user group to obtain the adjusted delivery effect contribution; The adjusted contribution of the delivery effect Among them, c it It represents the contribution of the i-th user group to the delivery effect in the t-th time period. T represents the total time span of the delivery effect data, that is, the number of time periods from the start of delivery to the current time point. α represents the time attenuation factor, 0<α≤1.

5. An advertisement placement information extraction system based on automated interaction, characterized in that: include: An interactive content acquisition module is used to acquire the interactive content between the user and the automated interactive system; A scheme identifier extraction module, used for extracting a scheme identifier according to the interactive content; A delivery plan determination module, used to match the delivery plan corresponding to the plan identifier in the stored database according to the plan identifier; A user information acquisition module, used to extract key information and recognize text from the interactive content to obtain user information; A delivery information set acquisition module is used to generate structured user portrait data according to the user information, and classify the user portrait data into corresponding delivery plans according to the delivery plans to obtain delivery information sets organized according to the delivery plans; An original data file acquisition module, used to acquire an original data file generated by an automated interactive system corresponding to the delivery plan, wherein the original data file includes historical interaction records with multiple users; A delivery effect related data extraction module is used to parse the original data file and extract delivery effect related data, wherein the delivery effect related data includes landing page information, interaction time information, IP address information, IP address type information, interaction entry source information and keyword information; An effect data statistics module, for counting effect data of the delivery scheme according to the delivery effect related data, wherein the effect data includes the number of chat conversations, whether user information is obtained, the amount of user information obtained, conversation efficiency and statistical date; An effect report generation module is used to generate an effect report of the delivery plan according to the effect data, and provide the effect report to the advertising delivery personnel for delivery effect analysis; The user group segmentation module is used to perform cluster analysis on data related to delivery effects and divide users into different user groups based on user characteristic parameters, including geographical distribution, interaction time distribution, and keyword preference; The user group delivery effect data acquisition module is used to collect statistics on the effect data of the corresponding delivery plans for different user groups and obtain the delivery effect data of each user group; The user group effect score calculation module is used to perform weighted summation on the delivery effect data of each user group to obtain the comprehensive effect score of each user group; The delivery effect contribution determination module is used to calculate the ratio of the comprehensive effect score of each user group to the total comprehensive effect scores of all user groups, and use the ratio as the delivery effect contribution of the corresponding user group; The key user group identification module is used to sort each user group according to the contribution of the delivery effect, and identify the key user group based on the sorting results; The key user group feature portrait adding module is used to summarize the feature portraits of key user groups according to the distribution rules of their user feature parameters, and add them to the effect report of the delivery plan.

6. The advertisement delivery information extraction system based on automated interaction according to claim 5, characterized in that: The scheme identification extraction module includes: A page link acquisition unit, used to acquire a page link in the interactive content; A link parameter acquisition unit, used to parse the page link to obtain page link parameters; A scheme identifier extraction unit is used to extract the scheme identifier from the page link parameter according to a preset extraction rule, wherein the extraction rule includes: It is determined in sequence according to a preset priority order whether there are parameters of a preset type in the page link parameters, the parameters of the preset type include numeric parameters and string parameters, and the numeric parameters and the string parameters can uniquely identify the delivery plan.

7. The advertisement delivery information extraction system based on automated interaction according to claim 5, characterized in that: The user information acquisition module includes: An information identification unit, used to identify key questions and user answers from the interactive content, wherein the key questions include name-related questions and contact information-related questions; A name information acquisition unit, used to analyze the user's answer to the name-related question according to a preset name judgment rule to obtain the user's name information; A contact information acquisition unit, configured to analyze the user's answer to the contact information-related question according to a preset contact information judgment rule to obtain the user's contact information; The region information acquisition unit is used to extract the interaction time information and IP address information from the interaction content, and perform region judgment according to the IP address information to obtain user region information.

Images