Network advertisement pushing intelligent management system based on natural language analysis processing

Abstract

The invention discloses a network advertisement pushing intelligent management system based on natural language analysis processing, and relates to the technical field of advertisement pushing management. Comprising a storage module which creates an advertisement database and is used for storing advertisement data to obtain a storage information set and extracting keyword information of the storage information set to obtain storage data items; and the acquisition module is used for acquiring natural language information input by a target user to obtain a natural information item, and obtaining a target advertisement item based on a matching relationship between the natural information item and the storage data item. According to the method, at least two pushing time nodes are set and correspond to browsing level items of different levels respectively, matching of the browsing level items is carried out according to the combination result of the real-time time items and the real-time chapter items, the target time threshold value is obtained through matching, and advertisement time is pushed according to the specific browsing time of the user. Advertisement pushing time can be changed according to specific demand conditions and interesting degrees of different users, and the intelligent effect of advertisement pushing is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of advertisement push management, and in particular to an intelligent management system for network advertisement push based on natural language analysis and processing. Background Art

[0002] Advertising management refers to the systematic management process of advertising, promotion and effect monitoring. Online advertising refers to the form of advertising promotion conducted through the Internet. The goal of advertising is determined, such as increasing brand awareness, increasing sales or attracting a specific customer base. Natural language processing is a sub-field of artificial intelligence that involves the interaction between computers and human language. It includes the analysis and understanding of natural language, enabling computers to understand, interpret, generate and respond to human language. In the process of advertising, it also needs to be combined with natural language analysis and processing to achieve more accurate advertising.

[0003] Patent publication number CN112330371A is an AI-based intelligent advertising push method and device, system and storage medium, which obtains screen information after the terminal is turned on; obtains program scene recognition results based on the screen information and the trained program scene recognition model; obtains corresponding advertisements based on the program scene recognition results; and plays the corresponding advertisements on the screen. Through the above methods, the present invention solves the problem that advertisements do not meet consumer needs and do not achieve the expected effect, and realizes intelligent advertising push.

[0004] When the above-mentioned and similar technical solutions push advertisements by acquiring the user's keyword information, different users have different mindsets when inputting keywords. For example, when a user inputs keywords with an attitude of understanding but not wanting to have a deep understanding, when advertisements are pushed based on the user's keyword analysis and processing, it is easy to cause the user to feel annoyed, resulting in poor results in intelligent advertisement push. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent management system for network advertisement push based on natural language analysis and processing to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above-mentioned purpose, the present invention provides the following technical solution: an intelligent management system for online advertising push based on natural language analysis and processing, comprising:

[0007] Storage module: create an advertisement database for storing advertisement data, obtain a storage information set, and extract keyword information of the storage information set to obtain storage data items;

[0008] Acquisition module: acquires the natural language information input by the target user, obtains the natural information item, and obtains the target advertisement item based on the matching relationship between the natural information item and the stored data item;

[0009] Analysis module: obtaining the browsing time and browsing length of the target user based on the state of the natural information item, obtaining the real-time time item and the real-time length item, setting the user's interest level to at least two levels, obtaining at least two browsing level items, and setting at least one push time node at the same time, obtaining at least two push time points, which correspond to browsing level items of different levels respectively, adjusting the push time through the first matching method, pushing advertisements to the target user, and adjusting the push quantity through the second matching method, pushing different quantities of advertisements to the target user;

[0010] The first matching method includes:

[0011] Step 1: Threshold setting: setting a time threshold and a text length threshold to obtain at least two time range items and at least two text length range items;

[0012] Step 2: Range combination: based on the combination results of the time range item and the length range item, at least four combination result items are obtained. Different combination result items represent different degrees of interest of the target users.

[0013] Step 3: setting a ratio, setting a division ratio, and segmenting the combined result item based on the division ratio to obtain at least two segmented result items, wherein the segmented result items correspond to the browsing level items respectively;

[0014] Step 4: Data matching, matching and combining ranges based on real-time time items and real-time length items, obtaining real-time combination items, and then obtaining target browsing level items, obtaining target push time nodes, and pushing advertisements to target users based on the target push time nodes.

[0015] Furthermore, the method for obtaining the time range item and the length range item includes: setting a first acquisition threshold, obtaining the average browsing time and the average browsing length of the target user based on the first acquisition threshold to obtain a first time item and an average length item, setting a cutting value, and obtaining at least one time cutting item and a length cutting item based on the combination of the average time item and the average length item with the cutting value, and then obtaining at least two time range items and at least two length range items.

[0016] Furthermore, the second matching method includes:

[0017] S1: data sorting, based on the results of sorting the target advertising items, to obtain the target sorting items;

[0018] S2: Matching value setting, setting at least two matching values, different matching values ​​correspond to different numbers of target advertisements in the target ranking items, and the matching values ​​correspond to the browsing level items respectively;

[0019] S3: Real-time matching value acquisition: based on the real-time time item and the real-time length item, range matching and combination results are performed to obtain the target matching value, and then the real-time matching value is obtained, and advertisements are pushed to the target users based on the real-time matching value.

[0020] Furthermore, the data sorting method includes: the number of keyword matches, based on the keyword features of the natural information items, obtaining the target feature set, obtaining the keyword features corresponding to the advertisements in the target advertising items, obtaining the corresponding feature set, and based on the sorting results of the corresponding feature set, placing the advertisement with the largest number of corresponding keyword features corresponding to the advertisements in the target advertising items at the top of the sorting results of the corresponding feature set, and so on, until the sorting results of the target advertising items are completed.

[0021] Furthermore, the method for obtaining the push time point includes:

[0022] M1: second threshold setting, setting a second acquisition threshold, acquiring the average browsing time of the target user based on the second acquisition threshold, and obtaining a second time item;

[0023] M2: Setting the first characteristic value: setting the interest characteristic value, obtaining the first target time value based on the combination result of the interest characteristic value and the second time item, and obtaining the push time point by taking the first target time value as the dividing point.

[0024] Furthermore, the method for obtaining the push time point includes:

[0025] N1: setting a third threshold value, setting a third acquisition threshold value, acquiring the number of views of the target user based on the third acquisition threshold value, and obtaining a set of the number of views;

[0026] N2: Time sorting, obtaining the browsing time corresponding to the browsing times in the browsing times set, and obtaining the time sorting items based on the sorting results of the browsing times;

[0027] N3: Setting the second feature value, setting the segmentation feature value, obtaining the target sorting item based on the combination of the segmentation feature value and the time sorting item, and obtaining the browsing time corresponding to the target sorting item to obtain the second target time value, and using the second target time value as the dividing point to obtain the push time point.

[0028] Furthermore, the method for obtaining the target advertisement item includes:

[0029] X1: Keyword extraction, based on the keyword extraction results of natural information items, obtain the target key terms;

[0030] X2: Keyword matching, based on the matching relationship between the target keyword and the stored data item, the target advertising item is obtained.

[0031] Furthermore, the method for acquiring the stored data item includes:

[0032] Z1: Topic acquisition, based on the storage information set, acquire the topic data of the stored advertisement data to obtain the topic data item;

[0033] Z2: key feature acquisition, based on the stored information set, the key feature data of the stored advertising data is acquired to obtain the key data items;

[0034] Z3: Data combination, the subject data item and the key data item correspond to the stored advertisement data respectively, and the stored data item is obtained based on the combination result of the subject data item and the key data item.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] The intelligent management system for network advertisement push based on natural language analysis and processing obtains natural language information input by users and keyword information in the natural information, matches corresponding advertisements in an advertisement content library, sets the browsing interest level to at least two levels, and sets at least two push time nodes, which correspond to browsing level items of different levels respectively. The browsing level items are matched according to the combination results of the real-time time item and the real-time length item, and the target time threshold is matched. The advertisement time is pushed according to the specific browsing time of the user. The advertisement push time can be changed according to the specific needs and interest levels of different users, thereby improving the intelligent effect of advertisement push.

[0037] At the same time, the target advertisements that match the keyword features in the natural language information input by the target user are sorted by a sorting method, and the advertisements are pushed in descending order according to the sorting results. The specific push time is set according to the average browsing time of the target user. This can make the push effect of the advertisement more closely related to the daily browsing habits of the target user, thereby improving the accuracy of the advertisement push, and the pushed advertisements are most closely related to the keyword features in the natural language information input by the target user.

[0038] Not only that, when pushing advertisements to target users, the number of advertisements pushed to target users is adjusted according to the target users' specific browsing time and browsing length. When it is determined that the target users are interested or very interested, the number of advertisements pushed is increased, thereby achieving the effect of selectively pushing different numbers of advertisements according to different user situations, further improving the intelligent effect of advertisement push and the advertisement push rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1A schematic diagram showing the relationship between the storage module, the acquisition module and the analysis module of the present invention;

[0040] Figure 2 This is a schematic diagram of the first matching method flow of the present invention;

[0041] Figure 3 This is a schematic diagram of the third push time and the first push time of the present invention;

[0042] Figure 4 It is a schematic diagram of the time range item of the present invention;

[0043] Figure 5 A schematic diagram showing the relationship between the number of views and the viewing time of the present invention;

[0044] Figure 6 Schematic diagram of segmentation feature values ​​of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] In the process of advertising push, when the user browses after searching for keywords, the differences in browsing time and browsing length indicate that the user has different interests in the keyword. When the user's browsing time is long and the browsing length is long, it means that the user's interest is relatively high at this time. At this time, the number of pushed advertisements can be appropriately increased. On the contrary, when the user's browsing time is short and the browsing length is also short, it means that the user's interest is not high. At this time, when advertising is pushed, it is easy to cause customers to feel bored and reduce the user's desire to continue browsing. When advertising is not pushed according to the user's interest level, not only the number of pushed advertisements is low, but also the quality of pushed advertisements will be poor. The technical solution provided by the present application first sets up an advertising content library, and then according to different The advertisements are stored based on keywords and topics, and the natural language information input by the user is obtained, and the keyword information in the natural information is obtained, which is used to match the corresponding advertisements in the advertisement content library to obtain the target advertisement item, and the browsing time and browsing length of the target user under the premise of the natural information are obtained to obtain the real-time time item and the real-time length item, and the browsing interest level is set to at least two levels to obtain the browsing level item, and at least two push time nodes are set, which correspond to the browsing level items of different levels respectively, and the browsing level items are matched according to the combination results of the real-time time item and the real-time length item, and the target time threshold is matched, and the target advertisement item is pushed according to the target time threshold, and the advertisement time is pushed according to the specific browsing time of the user, so as to improve the intelligent effect of advertisement push.

[0047] like Figure 1-Figure 3As shown, the present invention provides a technical solution: an intelligent management system for network advertising push based on natural language analysis and processing, including: a storage module: creating an advertising database for storing advertising data, obtaining a storage information set, and extracting keyword information of the storage information set to obtain a storage data item; an acquisition module: obtaining natural language information input by a target user, obtaining a natural information item, and obtaining a target advertising item based on a matching relationship between the natural information item and the storage data item; an analysis module: obtaining the browsing time and browsing length of the target user based on the state of the natural information item, obtaining a real-time time item and a real-time length item, setting the user's interest level to at least two levels, obtaining at least two browsing level items, and setting at least one push time node at the same time, obtaining at least two push time points, corresponding to browsing level items of different levels respectively, adjusting the push time through a first matching method, and pushing advertisements to the target user, through The push quantity is adjusted through the second matching method, and different numbers of advertisements are pushed to the target users; the first matching method includes: step one: threshold setting, setting the time threshold and the length threshold, and obtaining at least two time range items and at least two length range items; step two: range combination, based on the combination result of the time range item and the length range item, at least four combination result items are obtained, and different combination result items represent different degrees of interest of the target users; step three: ratio setting, setting the division ratio, segmenting the combination result item based on the division ratio, and obtaining at least two segmentation result items, and the segmentation result items correspond to the browsing level items respectively; step four: data matching, matching and combining the range based on the real-time time item and the real-time length item, obtaining the real-time combination item, and then obtaining the target browsing level item, obtaining the target push time node, and pushing advertisements to the target users based on the target push time node.

[0048] It should be noted that when creating the advertising database, use the MySQL relational database, which supports the SQL query language and is easy to implement data storage and retrieval. In the selected DBMS, create a new database, such as ad_database, and then create an "advertisement" table in the database, using the designed data model.

[0049] It should be noted that the user's interest level is set to three levels, namely, uninterested, interested and very interested, to obtain three browsing level items, and three push time nodes are set, namely, the first push time, the second push time and the third push time, wherein the third push time corresponds to uninterested, the second push time corresponds to interested, and the first push time corresponds to very interested. For example, the push time nodes are 20s, 30s and 40s, 20s is the first push time, 30s is the second push time, and 40s is the third push time. In the process of adjusting the push time, the time threshold and the length threshold are first set to obtain two time range items and two length range items, with the time threshold and the length threshold as nodes respectively. For example, the time threshold is 10s, and the length node is 100 words. At this time, the two time ranges are 0-10s and more than 10s respectively, and the two length ranges are 0-10s respectively. 0 words and more than 100 words, at this time, the combination results of the time range item and the length range item are four, by setting the division ratio, the division ratio is 25%, 50% and 25%, the four combination results are divided to obtain three segmentation result items, which correspond to the browsing level items respectively, wherein the segmentation result corresponding to not interested is the time range of 0-10s and the length range of 0-100 words, the segmentation result corresponding to interested is the time range of 0-10s and the length range of more than 100 words, the time range of more than 10s and the length range of 0-100 words, the segmentation result corresponding to very interested is the time range of more than 10s and the length range of more than 100 words, according to the matching and combination results of the real-time time item and the real-time length item, the corresponding segmentation results are matched, and then the target browsing level item is obtained, the target push time node is obtained, and the advertisement is pushed to the target user based on the target push time node.

[0050] It should be noted that when obtaining the user's interest level during browsing, the user's interest level directly affects the user's mood state. When the user is in a "not interested" state, the user will feel bored, indifferent or disappointed, which makes the user lack positive emotions when browsing, and may even feel irritated or repulsed. When the user is in an "interested" state, he will feel curious or slightly excited, have certain expectations for the content of the advertisement, and be willing to learn more about the relevant information. When the user is in a "very interested" state, the user's mood is usually more positive, and he will feel excited, expectant, and even enthusiastic. He is curious about the content of the advertisement and has a desire to explore. Therefore, the user's mood state can be fed back by obtaining the user's interest level, thereby better completing the push of advertisements.

[0051] like Figure 3As shown, in the specific implementation process, the real-time time item and real-time length item of the target user A are 5s and 50 words respectively. At this time, the browsing level corresponding to the segmented result of the target user obtained by matching is not interested, and the corresponding third push time is 40s. An advertisement is pushed to the target user in 40s. When the real-time time item and real-time length item of the target user B are 15s and 150 words respectively, the browsing level corresponding to the segmented result of the target user obtained by matching is very interested, and the corresponding first push time is 20s. An advertisement is pushed to the target user in 20s.

[0052] The method for obtaining time range items and length range items includes: setting a first acquisition threshold, obtaining the average browsing time and average browsing length of the target user based on the first acquisition threshold to obtain a first time item and an average length item, setting a cutting value, and obtaining at least one time cutting item and a length cutting item based on the combination of the average time item and the average length item with the cutting value, thereby obtaining at least two time range items and at least two length range items.

[0053] It should be noted that the first acquisition threshold is set to 48h. Within the first acquisition threshold, the average browsing time and the average browsing length of the target users are obtained, and the set cutting value is 50%. 50% of the average browsing time and the average browsing length of the target users are used as the time cutting item and the length cutting item, thereby obtaining two time range items and two length range items.

[0054] like Figure 4 As shown, in the specific implementation process, the average browsing time of the target users is 20s, and the average browsing length is 200 words. At this time, according to the set cutting values, the time cutting item and the length cutting item are 10s and 100 words respectively, and the two time range items and the two length range items are 0-10s, more than 10s and 0-100 words, more than 100 words respectively.

[0055] The second matching method includes: S1: data sorting, obtaining a target sorting item based on the result of sorting the target advertising item; S2: matching value setting, setting at least two matching values, different matching values ​​corresponding to different numbers of target advertisements in the target sorting item, and the matching values ​​corresponding to the browsing level items respectively; S3: real-time matching value acquisition, matching the range and combining the results based on the real-time time item and the real-time length item to obtain the target matching value, and then obtaining the real-time matching value, and pushing advertisements to the target users based on the real-time matching value.

[0056] It should be noted that the set matching values ​​are 1, 2 and 3, which correspond to the browsing level items respectively, that is, the matching value corresponding to not interested is 1, the matching value corresponding to interested is 2, and the matching value corresponding to very interested is 3. The range matching and combination results are performed based on the real-time time item and the real-time length item to obtain the target matching value, and then the real-time matching value is obtained, and advertisements are pushed to the target users based on the real-time matching value.

[0057] In the specific implementation process, the real-time time item and real-time length item of the target user are obtained as 17s and 140 words respectively. At this time, the browsing level corresponding to the segmented result of the target user is obtained through matching as very interested, and the corresponding first push time is 20s. An advertisement is pushed to the target user at 20s, and the real-time matching value is 3. At 20s, 3 advertisements are pushed to the target user.

[0058] The method for sorting data includes: the number of keyword matches, based on the keyword features of the natural information items, obtaining the target feature set, obtaining the keyword features corresponding to the advertisements in the target advertising items, obtaining the corresponding feature set, and based on the sorting results of the corresponding feature set, placing the advertisement with the largest number of corresponding keyword features corresponding to the advertisements in the target advertising items at the top of the sorting results of the corresponding feature set, and so on, until the sorting results of the target advertising items are completed.

[0059] It should be noted that the keyword features in the natural language information input by the target user are obtained, and the keyword features in the target advertising item are obtained at the same time. In order from high to low, the advertisement with the largest number of corresponding keyword features corresponding to the advertisements in the target advertising item is ranked first in the sorting result of the corresponding feature set.

[0060] In the specific implementation process, the keyword features obtained in the natural language information input by the target user are exercise and fat-reducing meals. At this time, the keywords corresponding to the advertisements in the target advertising items are exercise, fat-reducing meals, and the combination of exercise and fat-reducing meals. The advertisements in the target advertising items are sorted in order of the corresponding number of keyword features.

[0061] The method for obtaining the push time point includes: M1: setting the second threshold value, setting the second acquisition threshold, obtaining the average browsing time of the target user based on the second acquisition threshold, and obtaining the second time item; M2: setting the first feature value, setting the interest feature value, and obtaining the first target time value based on the combination result of the interest feature value and the second time item, and using the first target time value as the dividing point to obtain the push time point.

[0062] It should be noted that the second acquisition threshold is set at 48h, the average browsing time of the target users is obtained, and the interest feature values ​​are set at 40%, 60% and 80%. Based on the combination of the interest feature values ​​and the average browsing time of the target users, three first target time values ​​are obtained. The specific push time is set according to the average browsing time of the target users, which can make the advertising push effect closer to the daily browsing habits of the target users, thereby improving the accuracy of advertising push.

[0063] In the specific implementation process, the average browsing time of the target users is obtained to be 50 seconds. At this time, according to the set interest feature values, the three first target times are obtained to be 20 seconds, 30 seconds and 40 seconds respectively.

[0064] The method for obtaining the push time point includes: N1: setting the third threshold value, setting the third acquisition threshold, obtaining the number of browsing times of the target user based on the third acquisition threshold, and obtaining a browsing time set; N2: time sorting, obtaining the browsing time corresponding to the browsing time in the browsing time set, and obtaining the time sorting item based on the sorting result of the browsing time; N3: setting the second feature value, setting the segmentation feature value, obtaining the target sorting item based on the combination result of the segmentation feature value and the time sorting item, and obtaining the browsing time corresponding to the target sorting item to obtain the second target time value, and using the second target time value as the dividing point to obtain the push time point.

[0065] It should be noted that the third acquisition threshold is set to 24h, the number of browsing times of the target user is obtained, and the browsing time corresponding to different browsing times is sorted, the segmentation feature values ​​are set to 40%, 60% and 80%, and three second time values ​​are obtained based on the combination of the segmentation feature values ​​and the time sorting items. The push time point is selected after sorting the specific browsing time of the target user's different browsing times. When there are individual longer browsing times in a single browsing of the target user, this method can ensure as much as possible that the average browsing time of the target user will not be affected by the individual longer browsing time or the individual shorter browsing time.

[0066] like Figure 5-Figure 6 As shown, in the specific implementation process, the target user's browsing times are obtained to be 10 times, and the browsing times are 50s, 45s, 44s, 56s, 23s, 430s, 43s, 54s, 33s, and 20s respectively. According to the sorted results, the segmentation feature values ​​of 40%, 60% and 80% are selected as the second time values, which are 43s, 45s and 54s respectively.

[0067] The method for obtaining the target advertising item includes: X1: keyword extraction, obtaining the target keyword item based on the keyword extraction result of the natural information item; X2: keyword matching, obtaining the target advertising item based on the matching relationship between the target keyword item and the stored data item.

[0068] It should be noted that, based on the keyword features in the natural language information input by the target user, the keyword features in the stored data items are simultaneously obtained, and the target matched with the keyword features is used as the target advertisement item.

[0069] The method for acquiring stored data items includes: Z1: subject acquisition, based on the storage information set, acquiring the subject data of the stored advertising data to obtain the subject data item; Z2: key feature acquisition, based on the storage information set, acquiring the key feature data of the stored advertising data to obtain the key data item; Z3: data combination, the subject data item and the key data item correspond to the stored advertising data respectively, and the storage data item is obtained based on the combination result of the subject data item and the key data item.

[0070] It should be noted that when extracting keyword information from the storage information set, the keyword information not only includes subject information, but also includes key feature data during the display of advertising data. The subject information is combined with the key feature data to jointly extract the keyword information from the storage information set to obtain the storage data item.

[0071] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is limited by the attached embodiments and their equivalents.

Claims

1. An intelligent management system for online advertising push based on natural language analysis and processing, characterized in that: include: Storage module: create an advertisement database for storing advertisement data, obtain a storage information set, and extract keyword information of the storage information set to obtain storage data items; Acquisition module: acquires the natural language information input by the target user, obtains the natural information item, and obtains the target advertisement item based on the matching relationship between the natural information item and the stored data item; Analysis module: obtaining the browsing time and browsing length of the target user based on the state of the natural information item, obtaining the real-time time item and the real-time length item, setting the user's interest level to at least two levels, obtaining at least two browsing level items, and setting at least one push time node at the same time, obtaining at least two push time points, which correspond to browsing level items of different levels respectively, adjusting the push time through the first matching method, pushing advertisements to the target user, and adjusting the push quantity through the second matching method, pushing different quantities of advertisements to the target user; The first matching method includes: Step 1: Threshold setting: setting a time threshold and a text length threshold to obtain at least two time range items and at least two text length range items; Step 2: Range combination: based on the combination results of the time range item and the length range item, at least four combination result items are obtained. Different combination result items represent different degrees of interest of the target users. Step 3: setting a ratio, setting a division ratio, and segmenting the combined result item based on the division ratio to obtain at least two segmented result items, wherein the segmented result items correspond to the browsing level items respectively; Step 4: Data matching, matching and combining ranges based on real-time time items and real-time length items, obtaining real-time combination items, and then obtaining target browsing level items, obtaining target push time nodes, and pushing advertisements to target users based on the target push time nodes.

2. The intelligent management system for online advertising push based on natural language analysis and processing according to claim 1 is characterized by: The method for obtaining the time range item and the length range item includes: setting a first acquisition threshold, obtaining the average browsing time and the average browsing length of the target user based on the first acquisition threshold to obtain a first time item and an average length item, setting a cutting value, and obtaining at least one time cutting item and a length cutting item based on a combination of the average time item and the average length item with the cutting value, and then obtaining at least two time range items and at least two length range items.

3. The intelligent management system for online advertising push based on natural language analysis and processing according to claim 1 is characterized by: The second matching method includes: S1: data sorting, based on the results of sorting the target advertising items, to obtain the target sorting items; S2: Matching value setting, setting at least two matching values, different matching values ​​correspond to different numbers of target advertisements in the target ranking items, and the matching values ​​correspond to the browsing level items respectively; S3: Real-time matching value acquisition: based on the real-time time item and the real-time length item, range matching and combination results are performed to obtain the target matching value, and then the real-time matching value is obtained, and advertisements are pushed to the target users based on the real-time matching value.

4. The intelligent management system for online advertising push based on natural language analysis and processing according to claim 3 is characterized by: The data sorting method includes: obtaining a target feature set based on the keyword features of the natural information item according to the number of keyword matches, obtaining the keyword features corresponding to the advertisements in the target advertising item, obtaining a corresponding feature set, and based on the sorting result of the corresponding feature set, ranking the advertisement with the largest number of corresponding keyword features corresponding to the advertisements in the target advertising item as the first in the sorting result of the corresponding feature set, and so on, until the sorting result of the target advertising item is completed.

5. The intelligent management system for online advertising push based on natural language analysis and processing according to claim 1 is characterized by: The method for acquiring the push time point includes: M1: second threshold setting, setting a second acquisition threshold, acquiring the average browsing time of the target user based on the second acquisition threshold, and obtaining a second time item; M2: Setting the first characteristic value: setting the interest characteristic value, obtaining the first target time value based on the combination result of the interest characteristic value and the second time item, and obtaining the push time point by taking the first target time value as the dividing point.

6. The intelligent management system for online advertising push based on natural language analysis and processing according to claim 1 is characterized by: The method for acquiring the push time point includes: N1: setting a third threshold value, setting a third acquisition threshold value, acquiring the number of views of the target user based on the third acquisition threshold value, and obtaining a set of the number of views; N2: Time sorting, obtaining the browsing time corresponding to the browsing times in the browsing times set, and obtaining the time sorting items based on the sorting results of the browsing times; N3: Setting the second feature value, setting the segmentation feature value, obtaining the target sorting item based on the combination of the segmentation feature value and the time sorting item, and obtaining the browsing time corresponding to the target sorting item to obtain the second target time value, and using the second target time value as the dividing point to obtain the push time point.

7. The intelligent management system for online advertising push based on natural language analysis and processing according to claim 1 is characterized by: The method for obtaining the target advertisement item includes: X1: Keyword extraction, based on the keyword extraction results of natural information items, obtain the target key terms; X2: Keyword matching, based on the matching relationship between the target keyword and the stored data item, the target advertising item is obtained.

8. The intelligent management system for online advertising push based on natural language analysis and processing according to claim 1 is characterized by: The method for obtaining the stored data item includes: Z1: Topic acquisition, based on the storage information set, acquire the topic data of the stored advertisement data to obtain the topic data item; Z2: key feature acquisition, based on the stored information set, the key feature data of the stored advertising data is acquired to obtain the key data items; Z3: Data combination, the subject data item and the key data item correspond to the stored advertisement data respectively, and the stored data item is obtained based on the combination result of the subject data item and the key data item.

Images