An AI-based service platform data optimization system and method

Abstract

The present invention discloses a data optimization system and method for a service platform based on artificial intelligence, which relates to the technical field of service platform communication. It includes an access information collection module, an information matching module, a transmission information optimization module, and a data communication module. The access information collection module collects the access information of the artificial intelligence double-innovation service platform. After the user inputs a consultation question on the platform, the information matching module matches the reference data that can answer the corresponding input consultation question. The transmission information optimization module analyzes the marked situation of the matched reference data, and performs selective sorting processing on the matched reference data according to the marked situation. The data communication module generates the final optimized reference data and transmits the final optimized reference data to the client where the user who currently inputs the consultation question on the platform is located, improving the effectiveness of the data provided by the artificial intelligence double-innovation service platform and reducing the waste of data transmission resources at the same time.

Description

Technical Field

[0001] The present invention relates to the technical field of service platform communication, and specifically to a service platform data optimization system and method based on artificial intelligence. Background Art

[0002] An artificial intelligence dual-innovation service platform refers to a public platform that provides professional innovation and entrepreneurship services such as R & D tools, inspection and evaluation, security, standards, intellectual property rights, and entrepreneurship information consulting in the field of artificial intelligence. Users with information consulting needs can log in to the artificial intelligence dual-innovation service platform. After entering the questions they want to consult on the platform, the platform will transmit data related to solving the consulting questions to the user's client for the user to query and reference. Users can also provide information to the platform as information providers.

[0003] With the rapid growth of the amount of information stored on the artificial intelligence dual-innovation service platform, after a user enters a question they want to consult, the platform may provide a large amount of reference information, and the user needs to view the corresponding reference information one by one to find effective information. When receiving a large amount of reference information, the user may not be able to quickly find the truly effective data. Some users may exit the service platform after viewing some information and finding that there is no useful reference information for them. The existing service platforms cannot adaptively optimize the data providing method of the platform according to the user's information reception situation, reducing the information supply effectiveness of the artificial intelligence dual-innovation service platform. Summary of the Invention

[0004] The purpose of the present invention is to provide a service platform data optimization system and method based on artificial intelligence to solve the problems raised in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A service platform data optimization system based on artificial intelligence, the system includes an access information collection module, an information matching module, a transmission information optimization module, and a data communication module;

[0006] The access information collection module is used to collect the access information of the artificial intelligence dual-innovation service platform;

[0007] The information matching module is used to match reference data that can answer the corresponding input consulting questions after the user enters a consulting question on the platform;

[0008] The transmission information optimization module is used to analyze the marked situation of the matched reference data and perform selective sorting processing on the matched reference data according to the marked situation;

[0009] The data communication module is used to perform selective screening processing on the reference data after selective sorting processing, generate the final optimized reference data, and transmit the final optimized reference data to the client where the user who inputs a consultation question on the platform is located.

[0010] Preferably, the access information collection module includes a platform login data collection unit and an input information collection unit;

[0011] The platform login data collection unit is used to confirm the user's identity information after the user logs in to the artificial intelligence double-creation service platform, and the user's identity information is collected after the user grants permission;

[0012] The input information collection unit is used to collect the consultation question information input by the current user on the platform after logging in.

[0013] Preferably, the information matching module includes an input information recognition unit and a reference data retrieval unit;

[0014] The input information recognition unit is used to identify and perform semantic analysis on the consultation question information input by the user on the platform by using NLP technology;

[0015] The reference data retrieval unit is used to retrieve, from the database of the platform, the reference data that can answer the consultation question input by the user by using intelligent AI Q&A technology. The artificial intelligence double-creation service platform has an intelligent AI Q&A function. Intelligent AI Q&A technology is an innovative application based on artificial intelligence technology, and realizes the function of answering users' questions through natural language processing technology, that is, NLP technology and machine learning and other technologies.

[0016] Preferably, the transmission information optimization module includes a queried information statistics unit and a supply mode optimization unit;

[0017] The queried information statistics unit is used to count the feedback information after each item of reference data is viewed. After the user views each item of reference data, the platform will pop up the content of "whether the corresponding reference data is valid data" for the user to choose. If the user selects the "yes" option, the corresponding item of reference data will be marked 1 time; if the user selects the "no" option, the marking information will not be counted. The feedback information includes the marked times of each item of reference data that can answer the same consultation question, and there is more than one item of reference data that can answer the same consultation question;

[0018] The supply method optimization unit is used to, for the same consultation question, select whether to perform sorting processing on several reference data items according to the marked times of all reference data that can answer the corresponding consultation question: set a marked times difference threshold, and if the marked times difference degree of several reference data items exceeds the threshold, select to perform sorting processing on several reference data items; otherwise, do not perform sorting processing and confirm the final arrangement order of the reference data that can answer the same consultation question.

[0019] Preferably, the data communication module includes a supply information confirmation unit and a supply data transmission unit;

[0020] The supply information confirmation unit is used to collect the amount of reference data viewed by the user before exiting the platform due to the failure to find effective reference data in the past. When the user selects the "No" option before exiting the platform, it means that the user has not found effective reference data. Predict the acceptable data volume of the user based on the amount of reference data. If the user who currently inputs a consultation question is a user who logs in to the platform for the first time, do not perform screening processing on the reference data that can answer the consultation question input by the current user, and do not predict the acceptable data volume of the user who logs in to the platform for the first time. Confirm that the data provided to the current user is all the reference data that can answer the consultation question input by the current user; if the user who currently inputs a consultation question is not a user who logs in to the platform for the first time, perform screening processing on the reference data that can answer the consultation question input by the current user, and confirm that the data provided to the current user is the data after screening processing. Here, the reference data that can answer the consultation question input by the current user before screening processing is the reference data whose final arrangement order has been confirmed.

[0021] The supply data transmission unit is used to transmit the confirmed data provided to the current user to the client where the current user is located. The confirmed data provided to the current user is the final optimized reference data.

[0022] An artificial intelligence-based service platform data optimization method includes the following steps:

[0023] S1: Collect the access information of the artificial intelligence double-creation service platform;

[0024] S2: After the user inputs a consultation question on the platform, match the reference data that can answer the corresponding input consultation question;

[0025] S3: Analyze the marked situation of the matched reference data, and perform selective sorting processing on the matched reference data according to the marked situation;

[0026] S4: Perform selective screening processing on the reference data after selective sorting processing to generate the final optimized reference data, and transmit the final optimized reference data to the client where the user who currently inputs a consultation question on the platform is located.

[0027] Preferably, S1 includes: after the user logs in to the artificial intelligence mass entrepreneurship and innovation service platform, confirming the user's identity information and collecting the consultation question information input by the current user on the platform after logging in.

[0028] Preferably, S2 includes: using NLP technology to identify and semantically analyze the consultation question information input by the current user on the platform, and using intelligent AI question answering technology to retrieve all reference data from the platform's database that can answer the consultation question input by the current user.

[0029] Preferably, S3 includes: for all reference data that can answer the consultation question input by the current user, counting the marked information of each reference data after being viewed in the past, and the set of the historical marked times of each reference data is {H1, H2,... H n}, where n represents the number of reference data items that can answer the consultation question input by the current user. Calculate the marked times difference degree P of the n reference data according to the formula P = [∑ n i=1 (H i -(∑ n i=1 (H i )) / n) 2 / n] 1 / 2 , where i represents the i-th reference data. Set the marked times difference threshold as R, and compare P and R: if P > R, select to sort the n reference data, and sort the i-th reference data in descending order of the historical marked times. For reference data with the same marked times, use a random sorting method; if P ≤ R, do not perform sorting, that is, the n reference data are randomly arranged;

[0030] For the same consultation question, whether the reference data that can answer the corresponding consultation question is effective for different users, that is, whether it has reference value, cannot be accurately judged in the prior art. The present invention sets up a page pop-up on the service platform for the user to feedback whether the corresponding reference data is effective after viewing different reference data items. Compared with the prior art, the present invention can accurately obtain the user's feedback information on the reference data. Before transmitting the reference data that can answer the corresponding consultation question to the user, selective sorting is performed based on the marked situation of the reference data. The reason for performing selective sorting is to consider that there may be a situation where the difference in the marked times of the reference data is not significant. In this case, there is no need to perform sorting. Transmitting the sorted reference data to the user is beneficial to improving the probability that the user can quickly find effective reference data, and at the same time simplifies the unnecessary sorting work of the system.

[0031] Preferably, S4 includes: confirming whether the user who inputs the current consultation question is a user who logs in to the platform for the first time: if so, transmitting n items of reference data as the final optimized reference data to the client where the current user is located; if not, collecting the amount of reference data viewed by the current user before exiting the platform due to the failure to find effective reference data for the consulted question in the past: obtaining the reference data amount set as V = {V1, V2,... V m}, m represents the number of times the current user has exited the platform due to the failure to find effective reference data in the past, and predicting the acceptable data amount of the current user as V m+1 : V m+1 = θ * V m + (1 - θ) * D m , where D m represents the smoothed value of the amount of reference data viewed by the user before exiting the platform for the mth time, 0 < θ < 1, θ represents the smoothing coefficient, θ is the system default setting, obtaining that among the n items of reference data that have been selectively sorted, the total amount of the first k items of reference data does not exceed V m+1 and the total amount of the first k + 1 items of reference data exceeds V m+1 , and transmitting the first k items of reference data as the final optimized reference data to the client where the current user is located;

[0032] Optimize data transmission for different users: Considering that some users may log in to the service platform for the first time, while some users may have logged in several times. For users who log in to the service platform for the first time, since the historical viewing behavior data of the users cannot be obtained, all the reference data that can answer the corresponding consultation questions are directly transmitted to the client where the users are located; for users who have logged in to the platform, by analyzing the number of times the corresponding users have exited the platform due to the failure to find effective reference data in the past, the acceptable data amount of the users is predicted, that is, the users may exit the service platform after viewing a part of the data and finding no effective reference data. According to the acceptable data amount of the users, several items of reference data with appropriate data amounts are screened out, and the screened data is transmitted to the corresponding client where the users are located. Adaptive optimization of the data provision method of the platform according to the information reception situation of the users improves the information supply effectiveness of the artificial intelligence mass entrepreneurship and innovation service platform, and at the same time reduces the waste of data transmission resources.

[0033] Compared with the prior art, the beneficial effects of the present invention are:

[0034] Considering that for the same consulting question, it is impossible to accurately judge in the prior art whether the reference data that can answer the corresponding consulting question is valid for different users, that is, whether it has reference value. The present invention sets up a page pop-up on the service platform for users to feedback whether the corresponding reference data is valid after viewing different items of reference data. Compared with the prior art, the present invention can accurately obtain the feedback information of users on the reference data. Before transmitting the reference data that can answer the corresponding consulting question to the user, selective sorting processing is performed according to the marked situation of the reference data. The reason for performing selective sorting processing is that there may be a situation where the marked times of the reference data are not very different. In this case, it is not necessary to perform sorting processing. Transmitting the sorted reference data to the user is beneficial to improving the probability that the user can quickly find valid reference data, and at the same time simplifies the unnecessary sorting processing work of the system;

[0035] Data transmission optimization for different users: Considering that some users may log in to the service platform for the first time, while some users may have logged in several times. For users who log in to the service platform for the first time, since the historical viewing behavior data of the users cannot be obtained, all the reference data that can answer the corresponding consulting question is directly transmitted to the client where the user is located; for users who have logged in to the platform, by analyzing the number of times the corresponding users have exited the platform due to not finding valid reference data in the past, the acceptable data volume of the users is predicted, that is, the user may exit the service platform after viewing a part of the data and finding no valid reference data. According to the acceptable data volume of the users, several items of reference data with appropriate data volume are screened out, and the screened data is transmitted to the client where the corresponding user is located. The data providing method of the platform is adaptively optimized according to the information receiving situation of the users, which improves the information supply effectiveness of the artificial intelligence double-innovation service platform and reduces the waste of data transmission resources at the same time. Brief Description of the Drawings

[0036] Figure 1 It is a schematic structural diagram of a data optimization system of a service platform based on artificial intelligence according to the present invention;

[0037] Figure 2 It is a schematic flowchart of a data optimization method of a service platform based on artificial intelligence according to the present invention. Detailed Embodiments

[0038] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0039] Example 1: As Figure 1 shown, this embodiment provides a data optimization system for an artificial intelligence-based service platform. The system includes: an access information collection module, an information matching module, a transmission information optimization module, and a data communication module;

[0040] The access information collection module is used to collect access information of the artificial intelligence double-creation service platform;

[0041] The information matching module is used to match reference data that can answer the corresponding input consultation questions after the user enters a consultation question on the platform;

[0042] The transmission information optimization module is used to analyze the marked situation of the matched reference data and perform selective sorting processing on the matched reference data according to the marked situation;

[0043] The data communication module is used to perform selective screening processing on the reference data after selective sorting processing to generate the final optimized reference data, and transmit the final optimized reference data to the client where the user who entered the consultation question on the platform is located. Among them, the selective sorting processing refers to two situations: selecting to perform sorting processing on the data and selecting not to perform sorting processing on the data; the selective screening processing refers to two situations: selecting to perform screening processing on the data and selecting not to perform screening processing on the data.

[0044] The access information collection module includes a platform login data collection unit and an input information collection unit;

[0045] The platform login data collection unit is used to confirm the user's identity information after the user logs in to the artificial intelligence double-creation service platform;

[0046] The input information collection unit is used to collect the consultation question information entered by the current user on the platform after logging in.

[0047] The information matching module includes an input information recognition unit and a reference data retrieval unit;

[0048] The input information recognition unit is used to identify and perform semantic analysis on the consultation question information entered by the user on the platform using NLP technology;

[0049] The reference data retrieval unit is used to retrieve reference data that can answer the consultation questions entered by the user from the platform's database using intelligent AI question-answering technology.

[0050] The transmission information optimization module includes a queried information statistics unit and a supply mode optimization unit;

[0051] The queried information statistics unit is used to count the feedback information after each piece of reference data is viewed. After the user views each piece of reference data, the platform will pop up the content of "whether the corresponding reference data is valid data" for the user to select. If the user selects the "yes" option, the corresponding piece of reference data will be marked once; if the user selects the "no" option, the marking information will not be counted. The feedback information includes the number of times each piece of reference data that can answer the same consultation question is marked. There is more than one piece of reference data that can answer the same consultation question.

[0052] The supply method optimization unit is used to, for the same consultation question, based on the number of times each piece of reference data that can answer the corresponding consultation question is marked, select whether to sort several pieces of reference data: set the difference threshold of the number of marked times. If the difference degree of the number of marked times of several pieces of reference data exceeds the threshold, then select to sort several pieces of reference data; otherwise, do not perform sorting processing, and confirm the final arrangement order of the reference data that can answer the same consultation question.

[0053] The data communication module includes a supply information confirmation unit and a supply data transmission unit;

[0054] The supply information confirmation unit is used to collect the amount of reference data viewed by the user before exiting the platform due to not finding valid reference data in the past. If the user selects the "no" option before exiting the platform, it means that the user has not found valid reference data. Based on the amount of reference data, predict the acceptable data volume of the user. If the user who currently inputs the consultation question is a user who logs in to the platform for the first time, do not screen the reference data that can answer the consultation question input by the current user, and do not predict the acceptable data volume of the user who logs in to the platform for the first time. Confirm that the data provided to the current user is all the reference data that can answer the consultation question input by the current user; if the user who currently inputs the consultation question is not a user who logs in to the platform for the first time, screen the reference data that can answer the consultation question input by the current user, and confirm that the data provided to the current user is the data after screening. Here, the reference data that can answer the consultation question input by the current user before screening is the reference data whose final arrangement order has been confirmed.

[0055] The supply data transmission unit is used to transmit the confirmed data provided to the current user to the client where the current user is located. The confirmed data provided to the current user is the final optimized reference data.

[0056] Embodiment 2: As Figure 2 shown, this embodiment provides an artificial intelligence-based service platform data optimization method, which is implemented based on the service platform data optimization system in the embodiment, and specifically includes the following steps:

[0057] S1: Collect the access information of the artificial intelligence mass entrepreneurship and innovation service platform: After the user logs in to the artificial intelligence mass entrepreneurship and innovation service platform, confirm the user's identity information, and collect the consultation question information entered by the current user on the platform after logging in;

[0058] S2: Match the reference data that can answer the corresponding consultation questions entered by the user after the user enters the consultation questions on the platform: Use NLP technology to identify and perform semantic analysis on the consultation question information entered by the current user on the platform, and use intelligent AI question-answering technology to retrieve all the reference data that can answer the consultation questions entered by the current user from the platform's database;

[0059] S3: Analyze the marked situation of the matched reference data, and perform selective sorting processing on the matched reference data according to the marked situation: For all the reference data that can answer the consultation questions entered by the current user, count the marked information of each piece of reference data after being viewed in the past, and the set of the historical marked times of each piece of reference data is {H1, H2,... H n}, n represents the number of items of reference data that can answer the consultation questions entered by the current user. Calculate the marked times difference degree P of the n items of reference data according to the formula. P = [∑ n i=1 (H i -(∑ n i=1 (H i )) / n) 2 / n] 1 / 2 , i represents the i-th piece of reference data. Set the marked times difference threshold as R, and compare P and R: If P > R, select to perform sorting processing on the n items of reference data, and sort the i-th piece of reference data in descending order according to the historical marked times. The reference data with the same marked times is processed by random sorting; If P ≤ R, no sorting processing is performed, that is, the n items of reference data are randomly arranged;

[0060] S4: Perform selective screening processing on the reference data after selective sorting processing to generate the final optimized reference data, and transmit the final optimized reference data to the client where the user who entered the consultation questions on the platform is located: Confirm whether the user who entered the consultation questions currently is a user who logged in to the platform for the first time: If so, transmit the n items of reference data as the final optimized reference data to the current user's client; If not, collect the amount of reference data viewed by the current user before exiting the platform due to not finding effective reference data for the consultation questions in the past: Obtain the set of reference data amounts as V = {V1, V2,... V m}, m represents the number of times the current user exited the platform due to not finding effective reference data in the past. Predict the acceptable data amount of the current user as V m+1 : Vm+1 = θ * V m + (1 - θ) * D m where * represents the multiplication sign, and D m represents the smoothed value of the amount of reference data viewed by the user before the m-th exit from the platform. The smoothed value D1 of the amount of reference data viewed by the user before the 1st exit from the platform is obtained by solving D1 = θ * V1 + (1 - θ) * [(V1 + V2 + V3) / 3]. D2 is obtained by solving D2 = θ * V1 + (1 - θ) * D1, and D3 is obtained by solving D3 = θ * V2 + (1 - θ) * D2. By calculating in sequence, D is finally obtained m where 0 < θ < 1, θ represents the smoothing coefficient, and θ is the system default setting. It is obtained that among the n items of reference data that have undergone selective sorting, the total amount of data of the first k items of reference data does not exceed V m+1 and the total amount of data of the first k + 1 items of reference data exceeds V m+1 then the first k items of reference data are transmitted as the final optimized reference data to the client where the current user is located;

[0061] For example: It is predicted that the acceptable data volume of the current user is V m+1 = 10 KB. It is obtained that among the 10 items of reference data that have undergone selective sorting, the total amount of data of the first 5 items of reference data does not exceed V m+1 and the total amount of data of the first 6 items of reference data exceeds V m+1 then the first 6 items of reference data are transmitted as the final optimized reference data to the client where the current user is located.

[0062] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be regarded as limiting the claimed rights involved.

Claims

1. An artificial intelligence-based service platform data optimization system, characterized in that: The system includes an access information collection module, an information matching module, a transmission information optimization module, and a data communication module; The access information collection module is used to collect the access information of the artificial intelligence dual-innovation service platform; The information matching module is used to match the reference data that can answer the corresponding input consultation question after the user enters a consultation question on the platform; The transmission information optimization module is used to analyze the marked situation of the matched reference data, and perform selective sorting processing on the matched reference data according to the marked situation; The data communication module is used to perform selective screening processing on the reference data after the selective sorting processing, generate the final optimized reference data, and transmit the final optimized reference data to the client where the user who currently enters a consultation question on the platform is located; The transmission information optimization module includes a queried information statistics unit and a supply method optimization unit; The queried information statistics unit is used to count the feedback information after each item of reference data is viewed. After the user views each item of reference data, the platform will pop up the content of "whether the corresponding reference data is valid data" for the user to choose. If the user selects the "yes" option, the corresponding item of reference data is recorded as being marked once; if the user selects the "no" option, the marked information is not counted. The feedback information includes the marked times of each item of reference data that can answer the same consultation question; The supply method optimization unit is used to, for the same consultation question, select whether to perform sorting processing on several items of reference data according to the marked times of all the reference data that can answer the corresponding consultation question: set a marked times difference threshold. If the marked times difference degree of several items of reference data exceeds the threshold, select to perform sorting processing on several items of reference data; Otherwise, no sorting processing is performed, and the final arrangement order of the reference data that can answer the same consultation question is confirmed; The data communication module includes a supply information confirmation unit and a supply data transmission unit; The supply information confirmation unit is used to collect the amount of reference data viewed by the user before exiting the platform due to not finding valid reference data. If the user selects the "no" option before exiting the platform, it means that the user has not found valid reference data. Predict the acceptable data amount of the user according to the amount of reference data. If the user who currently enters a consultation question is a first-time user logging in to the platform, no screening processing is performed on the reference data that can answer the consultation question entered by the current user, and it is confirmed that the data provided to the current user is all the reference data that can answer the consultation question entered by the current user; If the user who currently enters a consultation question is not a first-time user logging in to the platform, screening processing is performed on the reference data that can answer the consultation question entered by the current user, and it is confirmed that the data provided to the current user is the data after the screening processing; The supply data transmission unit is used to transmit the confirmed data provided to the current user to the client where the current user is located.

2. The data optimization system of a service platform based on artificial intelligence according to claim 1, wherein: The access information collection module includes a platform login data collection unit and an input information collection unit; The platform login data collection unit is used to confirm the identity information of the user after the user logs in to the artificial intelligence dual-innovation service platform; The input information collection unit is used to collect the consultation question information input by the current user on the platform after logging in.

3. The data optimization system of an artificial intelligence-based service platform according to claim 2, wherein: The information matching module includes an input information recognition unit and a reference data retrieval unit; The input information recognition unit is used to identify and perform semantic analysis on the consultation question information input by the user on the platform by using NLP technology; The reference data retrieval unit is used to retrieve, from the database of the platform, reference data that can answer the consultation questions input by the user by using intelligent AI question-and-answer technology.

4. A method for optimizing service platform data based on artificial intelligence, applied to a system for optimizing service platform data based on artificial intelligence as described in any one of claims 1-3, characterized in that: It includes the following steps: S1: Collect the access information of the artificial intelligence mass entrepreneurship and innovation service platform; S2: After the user inputs a consultation question on the platform, match the reference data that can answer the corresponding input consultation question; S3: Analyze the marked situation of the matched reference data, and perform selective sorting processing on the matched reference data according to the marked situation; S4: Perform selective screening processing on the reference data after selective sorting processing to generate the final optimized reference data, and transmit the final optimized reference data to the client where the user who currently inputs the consultation question on the platform is located.

5. A data optimization method for an artificial intelligence-based service platform according to claim 4, characterized in that: S1 includes: After the user logs in to the artificial intelligence mass entrepreneurship and innovation service platform, confirm the user's identity information, and collect the consultation question information input by the current user on the platform after logging in.

6. The data optimization method of a service platform based on artificial intelligence according to claim 5, wherein: S2 includes: Use NLP technology to identify and perform semantic analysis on the consultation question information input by the current user on the platform, and use intelligent AI question-and-answer technology to retrieve all the reference data that can answer the consultation questions input by the current user from the database of the platform.

7. A method for optimizing service platform data based on artificial intelligence according to claim 6, characterized in that: The S3 includes: for all reference data that can answer the consultation questions input by the current user, count the marked information after each piece of reference data has been viewed in the past, and the set of the historical marked times of each piece of reference data is {H1, H2,... H n}}, where n represents the number of reference data items that can answer the consultation questions input by the current user. Calculate the difference degree P of the marked times of the n pieces of reference data according to the formula P = [∑ n i=1 (H i -(∑ n i=1 (H i )) / n) 2 / n] 1 / 2 , where i represents the i-th piece of reference data. Set the difference threshold of the marked times as R, and compare P and R: if P > R, select to sort the n pieces of reference data, and sort the i-th piece of reference data in descending order of the historical marked times. For the reference data with the same marked times, use the random sorting method; if P ≤ R, do not perform the sorting process.

8. A method for optimizing data of a service platform based on artificial intelligence according to claim 7, characterized in that: The S4 includes: confirming whether the user who currently inputs the consultation question is a user who logs in to the platform for the first time. If so, transmitting n items of reference data as the final optimized reference data to the client where the current user is located. If not, collecting the amount of reference data viewed by the current user before exiting the platform due to the failure to find effective reference data for the consulted question in the past. The obtained set of reference data amounts is V = {V1, V2,... V m}, m represents the number of times the current user has exited the platform due to the failure to find effective reference data in the past, and the acceptable data amount of the current user is predicted to be V m+1 : V m+1 = θ * V m + (1 - θ) * D m , where D m represents the smoothed value of the amount of reference data viewed by the user before the m-th exit from the platform, 0 < θ < 1, θ represents the smoothing coefficient. Obtaining that among the n items of reference data that have undergone selective sorting processing, the total amount of the first k items of reference data does not exceed V m+1 and the total amount of the first k + 1 items of reference data exceeds V m+1 , and transmitting the first k items of reference data as the final optimized reference data to the client where the current user is located.

Images