CN116701791B - Course recommendation method and system based on artificial intelligence - Google Patents

Course recommendation method and system based on artificial intelligence Download PDF

Info

Publication number
CN116701791B
CN116701791B CN202310895106.1A CN202310895106A CN116701791B CN 116701791 B CN116701791 B CN 116701791B CN 202310895106 A CN202310895106 A CN 202310895106A CN 116701791 B CN116701791 B CN 116701791B
Authority
CN
China
Prior art keywords
pages
page
browsing
webpage
click
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202310895106.1A
Other languages
Chinese (zh)
Other versions
CN116701791A (en
Inventor
郑楠
曹鹏宇
杨连增
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guoxin Blue Bridge Education Technology Co ltd
Original Assignee
Guoxin Blue Bridge Education Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guoxin Blue Bridge Education Technology Co ltd filed Critical Guoxin Blue Bridge Education Technology Co ltd
Priority to CN202310895106.1A priority Critical patent/CN116701791B/en
Publication of CN116701791A publication Critical patent/CN116701791A/en
Application granted granted Critical
Publication of CN116701791B publication Critical patent/CN116701791B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/955—Retrieval from the web using information identifiers, e.g. uniform resource locators [URL]
    • G06F16/9558—Details of hyperlinks; Management of linked annotations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/951—Indexing; Web crawling techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/284—Lexical analysis, e.g. tokenisation or collocates
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention discloses a course recommendation method and system based on artificial intelligence, which are characterized in that based on a browsing path, web pages with skip relations are clustered in a short time through three-dimensional clustering, web pages relevant to the browsing path are judged according to the browsing path and click keywords, a plurality of recommended courses are obtained through a course recommendation neural network, click texts of the web pages corresponding to the recommended courses are converted into pictures, course recommendation is carried out, a tree is established through the relationship of web page skip, a three-dimensional array is constructed according to the arrangement mode of the web pages in the tree, and the fact that the user always repeatedly skips the pages can be judged to extract courses which the user wants to purchase can be more accurately extracted. In the continuous switching of web pages, the web pages are placed at the conspicuous positions of the web pages, so that real-time and obvious recommendation is given to users in the browsing process.

Description

Course recommendation method and system based on artificial intelligence
Technical Field
The invention relates to the technical field of computers, in particular to a course recommendation method and system based on artificial intelligence.
Background
With the development of internet technology, many people learn and make questions on-line by selecting courses. Generally, after learning for a period of time, some test questions or challenging questions need to be made, missing knowledge points are found in the process of making questions, and then course learning is selected again in a targeted manner. For new users, it is necessary to recommend courses directly to the user.
At present, the problem to be solved is that a recommended course cannot be given to a user in real time in the process of changing web pages. In the prior art, a recommendation algorithm in a general case carries out user preference judgment through personal information and history information of a user to find corresponding data. However, the cost of purchasing courses by the user is high, if the user logs in the website initially, the course purchase or browsing of part of courses is not performed, the data is insufficient, and the recommendation can not be given by combining the browsing path of the user with keywords, so that the recommendation accuracy is low.
Disclosure of Invention
The invention aims to provide a course recommendation method and system based on artificial intelligence, which are used for solving the problems in the prior art.
In a first aspect, an embodiment of the present invention provides a method for recommending courses based on artificial intelligence, including:
obtaining a browsing path and a plurality of click texts; the click text is a text corresponding to the clicked webpage;
word segmentation is carried out on the click text to obtain a plurality of click keywords; the click keywords are words of click text of a user in a browsing path;
based on the browsing path, clustering web pages containing a jump relation in a preset time to obtain a first page set; the first page set comprises a plurality of webpage pages with the jump distance with the current page as a preset step length, and the preset step length is the preset jump times;
obtaining a second page set according to the browsing path and the click keywords; the second page set comprises repeated pages, associated non-browsed pages and keyword pages; the repeated page is a webpage page with the repeated browsing times being larger than the browsing times of other webpage pages in the browsing path; the associated non-browsed page indicates that the web page in the browsing path does not exist but can be browsed by two jumps, wherein the keyword page is a web page with the number of click keywords larger than the threshold value of the keyword;
obtaining a plurality of recommended courses according to the browsing path, the first page set and the second page set;
and recommending the recommended course to the user.
Optionally, based on the browsing path, clustering web pages that contain a skip relationship in a predetermined time to obtain a first page set, including:
establishing a three-dimensional array according to the browsing path and the webpage, and setting the position of the webpage in the browsing path in the three-dimensional array to be 1; the values in the three-dimensional array represent pages clicked by a user at different time points;
dividing the webpage in the browsing path according to time to obtain a historical browsing path and a current browsing page; the current browsing page is a webpage clicked by a current time point; the historical browsing path comprises all pages from a first webpage clicked by a user to a webpage clicked at a previous time point of the current time point;
three-dimensional clustering is carried out on the positions of the history browsing paths in the three-dimensional array to obtain a plurality of clustering clusters and a plurality of clustering center points;
according to the plurality of clustering center points, taking the cluster closest to the current time point as an adjacent cluster;
if the current browsed page is in the adjacent cluster, marking the webpage pages contained in the adjacent cluster as cluster pages, and setting the cluster pages into a first page set;
and if the current browsing page is not in the adjacent cluster, setting the current browsing page into the first page set.
Optionally, the establishing a three-dimensional array according to the browsing path and the web page, setting the position of the web page in the browsing path in the three-dimensional array to be 1, includes:
obtaining an initial page; the initial page is a main page of the website;
taking the initial page as a root node, and constructing a tree according to the jump condition of the webpage to obtain a webpage tree; child nodes in the webpage tree represent pages which can be skipped by the webpage pages corresponding to the parent nodes;
obtaining a fixed number; the fixed number is the number of the fixed clicked webpage pages;
taking the fixed number as the number of lines of a first dimension of the three-dimensional array, wherein the number of leaf nodes of the webpage tree is the number of columns of a second dimension, and the number of layers of the webpage tree is the number of pages of a third dimension;
and setting the position of the webpage corresponding to the click at the time point in the browsing path in the three-dimensional array as 1.
Optionally, the obtaining the second page set according to the browsing path and the click keyword includes:
obtaining a plurality of browsing pages and a plurality of corresponding browsing times according to the browsing paths;
according to the plurality of browsing pages, if the browsing times of the browsing pages are larger than the browsing times of other web pages, setting the browsing pages as repeated pages;
obtaining a plurality of associated non-browsed pages according to the plurality of browsed pages;
according to the plurality of click keywords, if the keyword pages are webpage pages with the number of the click keywords being larger than the keyword threshold value, setting the browsing pages as the keyword pages;
and setting the repeated page, a plurality of associated non-browsed pages and a plurality of keyword pages into an associated page set.
Optionally, the obtaining a plurality of associated non-browsed pages according to the plurality of browsed pages includes:
acquiring pages except the browsed page in the webpage page, and obtaining a plurality of unbrown pages;
constructing a directed graph of the webpage according to the jump relation of the webpage;
and obtaining the positions of the browsed pages in the directed graph, and finding out the unbrown pages which can be obtained by jumping with the browsed pages twice, so as to obtain a plurality of associated unbrown pages.
Optionally, the obtaining a plurality of recommended courses according to the browsing path, the first page set and the second page set includes:
inputting the browsing path into a long-term and short-term memory network to obtain browsing path output;
inputting the first page set and the second page set into a first neural network to obtain page output;
and outputting the browsing path and inputting the page output into a recommendation neural network to obtain a plurality of recommendation courses.
Optionally, the long-term and short-term memory network inputs the data in the browsing path from far to near according to time with step length of 1.
Optionally, recommending the recommended course to the user includes:
obtaining a recommended link and a recommended text according to the recommended course;
reconstructing a picture according to the recommended text to obtain a recommended picture; the recommended pictures comprise the content of texts corresponding to the recommended courses;
obtaining a dwell time and a replacement position; the stay time represents the stay time of a user on a webpage; the replacement position represents a position for recommending courses in the webpage;
and when the stay time is greater than a stay threshold value, replacing the picture at the replacement position with a recommended picture, and replacing the picture link at the replacement position with a recommended link.
Optionally, the obtaining a keyword page according to the multiple click keywords includes:
obtaining a target keyword according to the click keywords; the target keywords are keywords with the number of the click keywords being larger than that of other keywords;
and finding out a webpage corresponding to the target keyword to obtain a keyword page.
In a second aspect, embodiments of the present invention provide a system for artificial intelligence based course recommendation, comprising:
the acquisition module is used for: obtaining a browsing path and a plurality of click texts; the click text is a text corresponding to the clicked webpage;
keyword module: word segmentation is carried out on the click text to obtain a plurality of click keywords; the click keywords are words of click text of a user in a browsing path;
and a clustering module: based on the browsing path, clustering web pages containing a jump relation in a preset time to obtain a first page set; the first page set comprises a plurality of webpage pages with the jump distance with the current page as a preset step length, and the preset step length is the preset jump times;
repeating and associating module: obtaining a second page set according to the browsing path and the click keywords; the second page set comprises repeated pages, associated non-browsed pages and keyword pages; the repeated page is a webpage page with the repeated browsing times being larger than the browsing times of other webpage pages in the browsing path; the associated non-browsed page indicates that the web page in the browsing path does not exist but can be browsed by two jumps, wherein the keyword page is a web page with the number of click keywords larger than the threshold value of the keyword;
and a recommendation module: obtaining a plurality of recommended courses according to the browsing path, the first page set and the second page set;
and (3) a replacement module: and recommending the recommended course to the user.
Compared with the prior art, the embodiment of the invention achieves the following beneficial effects:
the embodiment of the invention also provides a course recommendation method and a course recommendation system based on artificial intelligence, wherein the method comprises the following steps: a browsing path and a plurality of click texts are obtained. The click text is a text corresponding to the clicked webpage. And segmenting the click text to obtain a plurality of click keywords. The click keyword is a word of a click text of a user in a browsing path. And clustering web pages containing a jump relation in a preset time based on the browsing path to obtain a first page set. The first page set comprises a plurality of webpage pages with the jump distance with the current page being a preset step length, and the preset step length is the preset jump times. And obtaining a second page set according to the browsing path and the clicking keywords. The second set of pages includes duplicate pages, associated unviewed pages, and keyword pages. The repeated page is a web page with the repeated browsing times larger than the browsing times of other web pages in the browsing path. The associated non-browsed page indicates that the web page in the browsing path does not exist but can be browsed by two jumps, and the keyword page is a web page with the number of click keywords larger than the keyword threshold value. And obtaining a plurality of recommended courses according to the browsing path, the first page set and the second page set. And recommending the recommended course to the user.
And building a tree through the relation of webpage page skip, and building a three-dimensional array according to the arrangement mode of the webpage pages in the tree and the time. The three-dimensional array can be used for reflecting the skip condition among the webpage pages when a user browses the webpage pages, namely whether to continuously skip or skip among a plurality of the webpage pages. Finding the page that the user is repeatedly jumping all the time can more accurately extract courses that the user wants to purchase. And finding out related webpage pages which are not browsed on the browsing path, the webpage pages with the largest repeated browsing times and the webpage pages corresponding to the keywords which are mentioned for a plurality of times, and extracting the webpage pages which are detected to be searched by the user but not seen by naked eyes or courses which are compared by the user all the time. And the judgment information is added, and the webpage pages and recommended courses which the user wants to find are extracted more accurately through the neural network. And the pictures are reconstructed according to courses, and the pictures are placed at the conspicuous positions of the web pages in the continuous switching of the web pages, so that the pictures are recommended to users in real time and obviously in the browsing process, and the accuracy of course recommendation is improved.
Drawings
FIG. 1 is a flow chart of a method for artificial intelligence based course recommendation provided by an embodiment of the present invention.
Fig. 2 is a schematic block diagram of an electronic device according to an embodiment of the present invention.
The marks in the figure: a bus 500; a receiver 501; a processor 502; a transmitter 503; a memory 504; bus interface 505.
Detailed Description
The present invention will be described in detail with reference to the accompanying drawings.
Example 1
As shown in FIG. 1, an embodiment of the present invention provides a method for artificial intelligence based course recommendation, the method comprising:
s101: obtaining a browsing path and a plurality of click texts; the click text is a text corresponding to the clicked webpage.
Wherein one click text corresponds to one webpage of the browsing path.
The browsing path is a mark of the webpage, which is recorded by the user clicking the webpage according to the time without breaking points. In the embodiment, numbers are adopted to label the webpage, and all the webpage of the recommended course are marked;
the browsing path is a clicked page in the time from the webpage opening of the user to the webpage clicked at the current time point.
S102: and segmenting the click text to obtain a plurality of click keywords. The click keyword is a word of a click text of a user in a browsing path.
Wherein the click keyword is a word related to information on a web page of a course desired by the user. For example, there are "machines" in the web page "
And the word segmentation algorithm is adopted to segment the click text. In this embodiment, the sentence piece model is used to segment words corresponding to the click text. The sentence piece model is a word segmentation tool based on a neural network.
S103: based on the browsing path, web pages with jump relations are clustered in a short time to obtain a first page set; the first page set comprises a plurality of webpage pages with the jump distance with the current page being a preset step length, and the preset step length is the preset jump times. The preset jump times are within a jump threshold. Specifically, the first page set may be obtained by three-dimensional clustering.
And taking the number of samples in the cluster as a jump threshold.
And the clustering page represents a webpage page of a clustering center point in a clustering cluster where the last time point is located.
In this embodiment, a k-means algorithm is used to perform three-dimensional clustering.
The reason why the last time point determination is adopted is that whether the current time point is in a cluster with the previous adjacent time point needs to be detected, but the cluster needs a lot of time points to determine.
S104: obtaining a second page set according to the browsing path and the click keywords; the second page set comprises repeated pages, associated non-browsed pages and keyword pages; the repeated page is a webpage page with the repeated browsing times being larger than the browsing times of other webpage pages in the browsing path; the associated non-browsed page indicates that the web page in the browsing path does not exist in the browsing path but can be jumped to the browsing path after two jumps; the keyword pages are web pages with the number of the clicked keywords larger than the keyword threshold value.
The method comprises the steps of displaying a webpage corresponding to a keyword with a large number of clicking times by a user.
Wherein the second set of pages is explained in a later method.
S105: and obtaining a plurality of recommended courses according to the browsing path, the first page set and the second page set.
Specifically, a plurality of recommended courses may be obtained through a course recommendation network, the course recommendation neural network including a long-short term memory network, a first neural network, and a recommendation neural network. Wherein, the artificial intelligence is to combine the user information and the browsing information and recommend through the neural network.
In this embodiment, the output number of the course recommendation neural network is 5, and then 5 recommendation numbers are obtained.
S106: and recommending the recommended course to the user.
After obtaining the plurality of recommended courses, the courses are recommended to the user. Specifically, the course can be directly set to the browsing page of the client, or the clicking text of the webpage corresponding to the recommended course can be converted into a picture to recommend the course.
Optionally, based on the browsing path, clustering web pages that contain a skip relationship in a predetermined time to obtain a first page set, including:
establishing a three-dimensional array according to the browsing path and the webpage, and setting the position of the webpage in the browsing path in the three-dimensional array to be 1; the values in the three-dimensional array represent pages that a user clicks on at different points in time.
The number of times that the user clicks on the webpage from far to near is marked as the row number of the array. Such as the time to click on a web page in the browsing path is 1 st second, 3 rd second, and 4 th second. The 1 st second is the first time of clicking on the webpage, and the line number of the three-dimensional array is set to 0. The 3 rd second is the second time of clicking the webpage, and the line number of the three-dimensional array is set to be 1. The 4 th second is the third time of clicking the webpage, and the line number of the three-dimensional array is set to be 2. The column number and page number of the three-dimensional array are marked by the jump condition of the webpage.
Dividing the webpage in the browsing path according to time to obtain a historical browsing path and a current browsing page; the current browsing page is a webpage clicked by a current time point; the history browsing path includes all pages from the first web page clicked by the user to the web page clicked at the last time point of the current time point.
And carrying out three-dimensional clustering on the positions of the history browsing paths in the three-dimensional array to obtain a plurality of clustering clusters and a plurality of clustering center points.
Wherein one cluster center point corresponds to one cluster.
Wherein the cluster represents the relevance of the webpage clicked by the user for a period of time. Because of the arrangement of the three-dimensional array, the pages in one cluster are short in the jumping process.
Wherein, a cluster center point corresponds to a webpage; one cluster includes a corresponding plurality of web pages at a plurality of points in time.
And taking the number of samples in the cluster as a jump threshold.
And taking the cluster closest to the current time point as the adjacent cluster according to the plurality of cluster center points.
And judging the difference value of the time corresponding to the current time point and the plurality of clustering centers, and taking the clustering cluster corresponding to the clustering center point with the difference value smaller than other difference values as the adjacent clustering cluster.
According to the situation that the clustering is carried out to obtain the frequent clustering, the webpage is set to be the webpage of the favorite query of the on-off time user;
if the current browsing page is in the adjacent cluster, marking the webpage page contained in the adjacent cluster as a clustered page, and setting the clustered page into a first page set.
And if the current browsing page is not in the adjacent cluster, setting the current browsing page to the first page set.
Through the method, the data are separated in a clustering mode. The user has a plurality of clusters adjacent to the path of the web page representation jump in the cluster. If the current browsed page of the user is in the adjacent cluster, the current browsed page is indicated to have a short jump path with the data clustered with the previous time point, and the user can think about courses to be purchased and interfaces to be jumped in the pages, so that related data are pushed to the user, and accuracy of course recommendation is improved.
Optionally, the establishing a three-dimensional array according to the browsing path and the web page, setting the position of the web page in the browsing path in the three-dimensional array to be 1, includes:
taking the initial page as a root node, and constructing a tree according to the jump condition of the webpage to obtain a webpage tree; and the child nodes in the webpage tree represent pages which can be jumped by the webpage pages corresponding to the parent nodes.
Wherein the web page tree is a shortest path tree.
Wherein, the web pages corresponding to the nodes with different layers can exist the same.
Obtaining a fixed number; the fixed number is the fixed number of clicked web pages.
The time length is fixed and is used for carrying out clustering judgment.
In this embodiment, the time length is set to be 50 times of clicking on the page by the user, and if the current time does not reach 50 times of clicking, zero padding is performed on the data earlier than the current time. If clicking only 7 times, the three-dimensional array with the number of rows of 50, with no data in the front 43 and data in the back 7, is constructed.
The fixed number is taken as the number of lines of the first dimension of the three-dimensional array, the number of leaf nodes of the webpage tree is the number of columns of the second dimension, and the number of layers of the webpage tree is the number of pages of the third dimension.
Wherein, the line numbers marked as an array from far to near by the time of clicking the webpage by the user. The different line numbers are only related to the order of the clicks, and are not related to the time interval of the clicks.
And marking the column numbers according to the jump relation by taking the initial page as 0. For example, in this embodiment, the initial page can jump to the first page and the fifth page, the first page can jump to the second page and the third page, and the fifth page can jump to the fourth page. And obtaining a webpage tree with the layer number of 3. Each layer has the same jump relation, representing jump links contained in a web page. The leaf nodes are a second page, a third page and a fourth page, and the number of the leaf nodes is 3. Therefore, in this embodiment, the three-dimensional array is 50×3x3.
And setting the correspondence of the webpage corresponding to the time point in the browsing path to 1 in the three-dimensional array.
Optionally, the obtaining the second page set according to the browsing path and the click keyword includes:
and obtaining a plurality of browsing pages and a plurality of corresponding browsing times according to the browsing paths.
The webpage page with the largest browsing times represents an array to be judged most.
And setting the browsing pages as repeated pages if the browsing times of the browsing pages are larger than the browsing times of other web pages according to the plurality of browsing pages.
And setting the repeated page as a page of the favorite query of the user.
And obtaining a plurality of associated non-browsed pages according to the plurality of browsed pages.
And setting the browsing page as a keyword page if the keyword page is a webpage page with the number of the click keywords larger than the keyword threshold value according to the plurality of click keywords.
In this embodiment, the keyword threshold is 3.
And setting the repeated page, a plurality of associated non-browsed pages and a plurality of keyword pages into an associated page set.
By the method, related webpage pages which are not browsed on the browsing path, webpage pages with the largest repeated browsing times and webpage pages corresponding to the keywords which are mentioned for a plurality of times are found and used for judging the webpage pages and recommended courses which the user wants to find.
Optionally, the obtaining a plurality of associated non-browsed pages according to the plurality of browsed pages includes:
acquiring pages except the browsed page in the webpage page, and obtaining a plurality of unbrown pages;
and constructing a directed graph of the webpage according to the jump relation of the webpage.
Wherein the directed graph represents a jump relationship between web pages.
In this embodiment, the web page is represented by numbers to obtain partial web pages such as {1,2,3}, where the page represented by 2 can be jumped to 3 and also to 1, so that the array of edges corresponding to the 2 vertices of the graph。
And obtaining the position of the browsed page in the directed graph, and finding the unbrown page which can be obtained by jumping with the browsed page twice, thereby obtaining the associated unbrown page.
By the method, the web page which is not browsed but has a short path can be accurately found, the web page which is wanted to be found by the user but not seen by naked eyes is mined, course recommendation is carried out on the basis of the web page, and accuracy of course recommendation is improved.
Optionally, the obtaining a plurality of recommended courses according to the browsing path, the first page set and the second page set includes:
and inputting the browsing path into a long-term and short-term memory network to obtain browsing path output.
The number of web pages set as the input LSTM (long-short-term memory network) in this embodiment is 15. The number of first layers of the LSTM is 15.
In this embodiment, the number of output browsing paths is 32.
And inputting the first page set and the second page set into a first neural network to obtain page output.
And outputting the browsing path and inputting the page output into a recommendation neural network to obtain a recommendation course.
The first neural network and the recommended neural network are deep neural networks with different structures.
By the method, the information browsed by the user is found according to the path, and the pages which the user probably wants to browse are obtained in the first page set and the second page set to predict courses which the user probably wants to search. The two judge recommended courses together, so that accuracy of course recommendation is improved.
Optionally, the long-term and short-term memory network inputs the data in the browsing path from far to near according to time with step length of 1.
In this embodiment, a fixed time length, that is, 50, is used to indicate that the user clicks the page 50 times. If the current time does not reach 50 clicks, zero padding is performed on the data earlier than the current time.
Inputting 50 data in the browsing path into a long-period and short-period memory network according to the time from far to near and the step length of 1;
the number of the first layers of the long-short-period memory network is the same as the fixed time length, which is 50 in this embodiment.
The last layer of the long-period memory network has the same number as the browsing path output.
Optionally, the converting the click text of the webpage corresponding to the recommended course into a picture, and performing course recommendation includes:
and obtaining a recommended link and a recommended text according to the recommended course.
The recommendation link is a link of a detailed page of the recommendation course, and the recommendation text is a text which is clicked into the recommendation course.
And reconstructing the picture according to the recommended text to obtain a recommended picture. The recommended pictures comprise the content of the text corresponding to the recommended courses.
The recommended pictures are pictures containing complete recommended texts.
Obtaining a dwell time and a replacement position; the stay time represents the stay time of a user on a webpage; the alternate location represents a location in the web page for recommending the course.
And when the stay time is greater than a stay threshold value, replacing the picture at the replacement position with a recommended picture, and replacing the picture link at the replacement position with a recommended link.
In this embodiment, the residence threshold is 10 seconds.
And converting the click text of the webpage corresponding to the recommended course into a picture, replacing the picture of the recommended course in the webpage, and recommending the course.
And a personalized interface is adopted for each user, so that the pictures displayed at the replacement positions are different, and the requirements of each user are met.
By the method, the picture is automatically built, the page at a certain fixed position is replaced to be the recommended course, and the combination of the webpage and the recommendation is carried out, so that the accuracy of course recommendation is improved.
Optionally, the obtaining a keyword page according to the multiple click keywords includes:
obtaining a target keyword according to the click keywords; the target keywords are keywords with the number of the click keywords being larger than that of other keywords;
and finding out a webpage corresponding to the target keyword to obtain a keyword page.
Example 2
Based on the course recommendation method based on the artificial intelligence, the embodiment of the invention also provides a course recommendation system based on the artificial intelligence, which comprises an acquisition module, a keyword module, a clustering module, a repetition and association module, a recommendation module and a replacement module.
The acquisition module is used for acquiring a browsing path and a plurality of click texts. The click text is a text corresponding to the clicked webpage.
The keyword module is used for segmenting the click text to obtain a plurality of click keywords. The click keyword is a word of a click text of a user in a browsing path.
The clustering module is used for clustering web pages containing a jump relation in preset time based on the browsing path to obtain a first page set. The first page set comprises a plurality of webpage pages with the jump distance with the current page being a preset step length, and the preset step length is the preset jump times.
The repeated and associated module is used for obtaining a second page set according to the browsing path and the click keyword; the second page set comprises repeated pages, associated non-browsed pages and keyword pages; the repeated page is a webpage page with the repeated browsing times being larger than the browsing times of other webpage pages in the browsing path; the associated non-browsed page indicates that the web page in the browsing path does not exist but can be browsed by two jumps, and the keyword page is a web page with the number of click keywords larger than the keyword threshold value.
The recommendation module is used for obtaining a plurality of recommendation courses according to the browsing path, the first page set and the second page set.
The replacement module is used for recommending the recommended course to the user.
An embodiment of the present invention further provides an electronic device, as shown in fig. 2, including a memory 504, a processor 502, and a computer program stored in the memory 504 and executable on the processor 502, where the processor 502 implements the steps of any of the methods of course recommendation based on artificial intelligence described above when executing the program.
Where in FIG. 2 a bus architecture (represented by bus 500), bus 500 may include any number of interconnected buses and bridges, with bus 500 linking together various circuits, including one or more processors, represented by processor 502, and memory, represented by memory 504. Bus 500 may also link together various other circuits such as peripheral devices, voltage regulators, power management circuits, etc., as are well known in the art and, therefore, will not be described further herein. Bus interface 505 provides an interface between bus 500 and receiver 501 and transmitter 503. The receiver 501 and the transmitter 503 may be the same element, i.e. a transceiver, providing a means for communicating with various other apparatus over a transmission medium. The processor 502 is responsible for managing the bus 500 and general processing, while the memory 504 may be used to store data used by the processor 502 in performing operations.
Embodiments of the present invention also provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of any of the methods of artificial intelligence based course recommendation described above, as well as the data referred to above.

Claims (8)

1. A method for artificial intelligence based course recommendation, comprising:
obtaining a browsing path and a plurality of click texts; the click text is a text corresponding to the clicked webpage;
word segmentation is carried out on the click text to obtain a plurality of click keywords; the click keywords are words of click text of a user in a browsing path;
based on the browsing path, clustering web pages containing a jump relation in a preset time to obtain a first page set; the first page set comprises a plurality of webpage pages with the jump distance with the current page as a preset step length, and the preset step length is the preset jump times;
obtaining a second page set according to the browsing path and the click keywords; the second page set comprises repeated pages, associated non-browsed pages and keyword pages; the repeated page is a webpage page with the repeated browsing times being larger than the browsing times of other webpage pages in the browsing path; the associated non-browsed page indicates that the web page in the browsing path does not exist in the browsing path but can be jumped to the browsing path after two jumps; the keyword pages are web pages with the number of the clicked keywords larger than the keyword threshold value;
obtaining a plurality of recommended courses according to the browsing path, the first page set and the second page set;
recommending the recommended course to the user;
based on the browsing path, clustering web pages containing a jump relation in a preset time to obtain a first page set, which comprises the following steps:
establishing a three-dimensional array according to the browsing path and the webpage, and setting the position of the webpage in the browsing path in the three-dimensional array to be 1; the values in the three-dimensional array represent pages clicked by a user at different time points;
dividing the webpage in the browsing path according to time to obtain a historical browsing path and a current browsing page; the current browsing page is a webpage clicked by a current time point; the historical browsing path comprises all pages from a first webpage clicked by a user to a webpage clicked at a previous time point of the current time point;
three-dimensional clustering is carried out on the positions of the history browsing paths in the three-dimensional array to obtain a plurality of clustering clusters and a plurality of clustering center points;
according to the plurality of clustering center points, taking the cluster closest to the current time point as an adjacent cluster;
if the current browsed page is in the adjacent cluster, marking the webpage pages contained in the adjacent cluster as cluster pages, and setting the cluster pages into a first page set;
if the current browsing page is not in the adjacent cluster, setting the current browsing page into the first page set;
the establishing a three-dimensional array according to the browsing path and the web page, setting the position of the web page in the browsing path in the three-dimensional array to be 1, includes:
obtaining an initial page; the initial page is a main page of the website;
taking the initial page as a root node, and constructing a tree according to the jump condition of the webpage to obtain a webpage tree; child nodes in the webpage tree represent pages which can be skipped by the webpage pages corresponding to the parent nodes;
obtaining a fixed number; the fixed number is the number of the fixed clicked webpage pages;
taking the fixed number as the number of lines of a first dimension of the three-dimensional array, wherein the number of leaf nodes of the webpage tree is the number of columns of a second dimension, and the number of layers of the webpage tree is the number of pages of a third dimension;
and setting the position of the webpage corresponding to the click at the time point in the browsing path in the three-dimensional array as 1.
2. The method for artificial intelligence based course recommendation of claim 1, wherein the obtaining a second set of pages from the browsing path and the click keyword comprises:
obtaining a plurality of browsing pages and a plurality of corresponding browsing times according to the browsing paths;
according to the plurality of browsing pages, if the browsing times of the browsing pages are larger than the browsing times of other web pages, setting the browsing pages as repeated pages;
obtaining a plurality of associated non-browsed pages according to the plurality of browsed pages;
according to the plurality of click keywords, if the keyword pages are webpage pages with the number of the click keywords being larger than the keyword threshold value, setting the browsing pages as the keyword pages;
and setting the repeated page, a plurality of associated non-browsed pages and a plurality of keyword pages into an associated page set.
3. The method for artificial intelligence based course recommendation of claim 1, wherein the obtaining a second set of pages from the browsing path and the click keyword comprises:
obtaining a plurality of browsing pages and a plurality of corresponding browsing times according to the browsing paths;
according to the plurality of browsing pages, if the browsing times of the browsing pages are larger than the browsing times of other web pages, setting the browsing pages as repeated pages;
obtaining a plurality of associated non-browsed pages according to the plurality of browsed pages;
according to the plurality of click keywords, if the keyword pages are webpage pages with the number of the click keywords being larger than the keyword threshold value, setting the browsing pages as the keyword pages;
and setting the repeated page, a plurality of associated non-browsed pages and a plurality of keyword pages into an associated page set.
4. The method for artificial intelligence based course recommendation of claim 1, wherein the obtaining a plurality of recommended courses based on the browsing path, the first page set and the second page set comprises:
inputting the browsing path into a long-term and short-term memory network to obtain browsing path output;
inputting the first page set and the second page set into a first neural network to obtain page output;
and outputting the browsing path and inputting the page output into a recommendation neural network to obtain a plurality of recommendation courses.
5. The artificial intelligence based course recommendation method of claim 4 wherein the long and short term memory network is to input data in a browsing path in steps of 1 from far to near in time.
6. The artificial intelligence based course recommendation method of claim 1, wherein recommending recommended courses to a user comprises:
obtaining a recommended link and a recommended text according to the recommended course;
reconstructing a picture according to the recommended text to obtain a recommended picture; the recommended pictures comprise the content of texts corresponding to the recommended courses;
obtaining a dwell time and a replacement position; the stay time represents the stay time of a user on a webpage; the replacement position represents a position for recommending courses in the webpage;
and when the stay time is greater than a stay threshold value, replacing the picture at the replacement position with a recommended picture, and replacing the picture link at the replacement position with a recommended link.
7. The method for artificial intelligence based course recommendation of claim 1, wherein obtaining a keyword page from the plurality of click keywords comprises:
obtaining a target keyword according to the click keywords; the target keywords are keywords with the number of the click keywords being larger than that of other keywords;
and finding out a webpage corresponding to the target keyword to obtain a keyword page.
8. A system for artificial intelligence based course recommendation, comprising:
the acquisition module is used for: obtaining a browsing path and a plurality of click texts; the click text is a text corresponding to the clicked webpage;
keyword module: word segmentation is carried out on the click text to obtain a plurality of click keywords; the click keywords are words of click text of a user in a browsing path;
and a clustering module: based on the browsing path, clustering web pages containing a jump relation in a preset time to obtain a first page set; the first page set comprises a plurality of webpage pages with the jump distance with the current page as a preset step length, and the preset step length is the preset jump times;
repeating and associating module: obtaining a second page set according to the browsing path and the click keywords; the second page set comprises repeated pages, associated non-browsed pages and keyword pages; the repeated page is a webpage page with the repeated browsing times being larger than the browsing times of other webpage pages in the browsing path; the associated non-browsed page indicates that the web page in the browsing path does not exist but can be browsed by two jumps, wherein the keyword page is a web page with the number of click keywords larger than the threshold value of the keyword;
and a recommendation module: obtaining a plurality of recommended courses according to the browsing path, the first page set and the second page set;
and (3) a replacement module: recommending the recommended course to the user;
based on the browsing path, clustering web pages containing a jump relation in a preset time to obtain a first page set, which comprises the following steps:
establishing a three-dimensional array according to the browsing path and the webpage, and setting the position of the webpage in the browsing path in the three-dimensional array to be 1; the values in the three-dimensional array represent pages clicked by a user at different time points;
dividing the webpage in the browsing path according to time to obtain a historical browsing path and a current browsing page; the current browsing page is a webpage clicked by a current time point; the historical browsing path comprises all pages from a first webpage clicked by a user to a webpage clicked at a previous time point of the current time point;
three-dimensional clustering is carried out on the positions of the history browsing paths in the three-dimensional array to obtain a plurality of clustering clusters and a plurality of clustering center points;
according to the plurality of clustering center points, taking the cluster closest to the current time point as an adjacent cluster;
if the current browsed page is in the adjacent cluster, marking the webpage pages contained in the adjacent cluster as cluster pages, and setting the cluster pages into a first page set;
if the current browsing page is not in the adjacent cluster, setting the current browsing page into the first page set;
the establishing a three-dimensional array according to the browsing path and the web page, setting the position of the web page in the browsing path in the three-dimensional array to be 1, includes:
obtaining an initial page; the initial page is a main page of the website;
taking the initial page as a root node, and constructing a tree according to the jump condition of the webpage to obtain a webpage tree; child nodes in the webpage tree represent pages which can be skipped by the webpage pages corresponding to the parent nodes;
obtaining a fixed number; the fixed number is the number of the fixed clicked webpage pages;
taking the fixed number as the number of lines of a first dimension of the three-dimensional array, wherein the number of leaf nodes of the webpage tree is the number of columns of a second dimension, and the number of layers of the webpage tree is the number of pages of a third dimension;
and setting the position of the webpage corresponding to the click at the time point in the browsing path in the three-dimensional array as 1.
CN202310895106.1A 2023-07-20 2023-07-20 Course recommendation method and system based on artificial intelligence Active CN116701791B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310895106.1A CN116701791B (en) 2023-07-20 2023-07-20 Course recommendation method and system based on artificial intelligence

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310895106.1A CN116701791B (en) 2023-07-20 2023-07-20 Course recommendation method and system based on artificial intelligence

Publications (2)

Publication Number Publication Date
CN116701791A CN116701791A (en) 2023-09-05
CN116701791B true CN116701791B (en) 2023-10-20

Family

ID=87835976

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310895106.1A Active CN116701791B (en) 2023-07-20 2023-07-20 Course recommendation method and system based on artificial intelligence

Country Status (1)

Country Link
CN (1) CN116701791B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117575745B (en) * 2024-01-17 2024-04-30 山东正禾大教育科技有限公司 Course teaching resource individual recommendation method based on AI big data

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101599089A (en) * 2009-07-17 2009-12-09 中国科学技术大学 The automatic search of update information on content of video service website and extraction system and method
CN105912604A (en) * 2016-04-05 2016-08-31 苏州奇展信息科技有限公司 On-line training platform for customized recommendation of curriculums
KR20200003665A (en) * 2018-07-02 2020-01-10 서울시립대학교 산학협력단 Method and server for recommending webpage
CN113591003A (en) * 2021-07-30 2021-11-02 平安普惠企业管理有限公司 Webpage processing method and device, computer equipment and storage medium
WO2023279595A1 (en) * 2021-07-07 2023-01-12 苏州浪潮智能科技有限公司 Method and apparatus for optimizing b/s architecture page response time, device, and medium
CN115705291A (en) * 2021-08-11 2023-02-17 腾讯科技（深圳）有限公司 Data processing method, device and computer readable storage medium

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101599089A (en) * 2009-07-17 2009-12-09 中国科学技术大学 The automatic search of update information on content of video service website and extraction system and method
CN105912604A (en) * 2016-04-05 2016-08-31 苏州奇展信息科技有限公司 On-line training platform for customized recommendation of curriculums
KR20200003665A (en) * 2018-07-02 2020-01-10 서울시립대학교 산학협력단 Method and server for recommending webpage
WO2023279595A1 (en) * 2021-07-07 2023-01-12 苏州浪潮智能科技有限公司 Method and apparatus for optimizing b/s architecture page response time, device, and medium
CN113591003A (en) * 2021-07-30 2021-11-02 平安普惠企业管理有限公司 Webpage processing method and device, computer equipment and storage medium
CN115705291A (en) * 2021-08-11 2023-02-17 腾讯科技（深圳）有限公司 Data processing method, device and computer readable storage medium

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
"一个基于web访问路径聚类的智能推荐系统";宋江春等;《信息与控制》;参照第1-4节 *
"基于动态矩阵预测的网页实时推荐研究";亓俊红等;《微计算机应用》;第760-763页 *
Shunlai Bai等."Search Results Clustering Based on Suffix Array and VSM".《 2010 IEEE/ACM Int'l Conference on Green Computing and Communications & Int'l Conference on Cyber, Physical and Social Computing》.2011,第1-4页. *

Also Published As

Publication number Publication date
CN116701791A (en) 2023-09-05

Similar Documents

Publication Publication Date Title
CN110046304B (en) User recommendation method and device
CN111444428B (en) Information recommendation method and device based on artificial intelligence, electronic equipment and storage medium
US20230146960A1 (en) Temporally Sequenced Content Recommender Method and System
CN108829822B (en) Media content recommendation method and device, storage medium and electronic device
CN1934569B (en) Search systems and methods with integration of user annotations
CN111382361B (en) Information pushing method, device, storage medium and computer equipment
CN109726274B (en) Question generation method, device and storage medium
CN109190049B (en) Keyword recommendation method, system, electronic device and computer readable medium
CN111859160B (en) Session sequence recommendation method and system based on graph neural network
CN104268292B (en) The label Word library updating method of portrait system
US20150095278A1 (en) Adaptive Probabilistic Semantic System and Method
CN104063476A (en) Social network-based content recommending method and system
Strobbe et al. Interest based selection of user generated content for rich communication services
CN111625715B (en) Information extraction method and device, electronic equipment and storage medium
CN116701791B (en) Course recommendation method and system based on artificial intelligence
CN108205572A (en) A kind of searching method, device and equipment
CN110765348B (en) Hot word recommendation method and device, electronic equipment and storage medium
CN113704420A (en) Method and device for identifying role in text, electronic equipment and storage medium
CN108153754B (en) Data processing method and device
CN111489196B (en) Prediction method and device based on deep learning network, electronic equipment and medium
CN116956183A (en) Multimedia resource recommendation method, model training method, device and storage medium
CN116541486A (en) News information aggregation method based on data mining and deep learning
Han et al. Developing smart service concepts: morphological analysis using a Novelty-Quality map
Ozsoy Utilizing fasttext for venue recommendation
CN114580533A (en) Method, apparatus, device, medium, and program product for training feature extraction model

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant