A book recommendation method and system based on artificial intelligence

Abstract

The present application relates to the field of data processing, and in particular to a method and system for recommending books based on artificial intelligence, the method comprising: classifying articles and segmenting words to obtain words and the categories to which the words belong, and calculating the importance of the words; generating a recommendation list according to the importance of the words, obtaining the reading time of the articles in the recommendation list that the user browses; dividing the reading time by the number of words in the article as the user's reading speed; calculating the user's interest in the recommendation list; obtaining word-related evaluations based on the interest level, and updating the recommendation list based on the word-related evaluations in response to a refresh instruction. The present application can improve the accuracy of book recommendations and better meet the needs of users.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to an artificial intelligence-based book recommendation method and system. Background Art

[0002] The article search platform can intelligently classify articles by subject and field through intelligent algorithms. This classification is not only based on the keywords and abstracts of the articles, but also may use machine learning models to analyze the full text content to ensure the accuracy of the classification. Users can choose the corresponding classification for in-depth reading based on their professional background, research interests or current knowledge needs.

[0003] However, the recommendation method of the existing technology is relatively simple, and there is no connection between each recommendation. It is impossible to achieve the effect that the recommended articles are more in line with the user's needs as the user uses the platform for a longer time. Summary of the invention

[0004] In order to make the recommended book articles more accurately meet the needs of users, the present application provides a book recommendation method and system based on artificial intelligence.

[0005] In the first aspect, the present application provides a book recommendation method based on artificial intelligence, which adopts the following technical solution:

[0006] A book recommendation method based on artificial intelligence comprises the following steps: classifying articles and segmenting words to obtain words and the categories to which the words belong, and calculating the importance of words; generating a recommendation list according to the importance of words, and obtaining the reading time of the articles in the recommendation list browsed by the user; dividing the reading time by the number of words in the article as the user's reading speed; calculating the user's interest in the recommendation list; obtaining word-related evaluations according to the interest level, and updating the recommendation list according to the word-related evaluations in response to a refresh instruction; wherein the method for calculating the interest level comprises: classifying the user's reading speed by time to obtain a reading speed set; using a preset detection model to detect and delete abnormal data in the reading speed set to obtain a reading habit set; taking the average of the reading speeds in the reading habit set as the user's reading ability; calculating the interest level, and the calculation formula is:

[0007] ,in, is the user's interest level, For the user's reading speed, For the user's reading ability, For The exponential function with base , represents the minimum function, represents the maximum value function, represents a set of reading habits, Represents the coefficient parameter.

[0008] Optionally, calculating the word importance includes the steps of: calculating the first prominence of each word in each category, the first prominence being the ratio of the number of occurrences of each word to the total number of all words in the category to which each word belongs; calculating the second prominence of each word in the article, the calculation formula being: ,in, is the second prominence of the word, is the starting point of accumulation, is the accumulation end point, For words in category The first prominence degree of the word; the second prominence degree is taken as the word importance degree.

[0009] Optionally, the first prominence of each word in each category is calculated, where the first prominence is the ratio of the number of occurrences of each word to the total number of all words in the category to which each word belongs; the second prominence of each word in the article is calculated using the following formula: ,in, is the second prominence of the word, is the starting point of accumulation, is the accumulation end point, For words in category The first prominence degree of the word; modify the second prominence degree to obtain the word importance.

[0010] Optionally, the correction process includes the steps of: sorting the words according to the order of the article to obtain the text; clustering the positions of the words in the text according to the clustering algorithm to obtain the maximum cluster of the words, one word corresponds to one maximum cluster, wherein a word is set as the target word, the input of the clustering is the position of the target word, the output of the clustering is a number of clusters, each cluster contains multiple target words, the cluster with the largest number of target words is taken as the maximum cluster, and the number of target words is the number of samples; the calculation formula of the word importance is: ,in, is the importance of the word, The second most prominent degree of the word, is the number of samples in the largest cluster corresponding to the word, For An exponential function with base .

[0011] Optionally, the article is classified and the words are segmented to obtain the words and the categories to which the words belong, including the steps of: classifying the article into categories, the categories including: title, abstract, chapter title, main text and references; using a word segmentation tool, segmenting the text into words to obtain independent words after segmentation, and determining the category to which the word belongs.

[0012] Optionally, according to the interest level, a word relevance evaluation is obtained, including: taking the product of the interest level of the user's read article and the word importance as the initial word relevance evaluation, and taking the average of multiple initial word relevance evaluations as the word relevance evaluation.

[0013] Optionally, abnormal data includes: abnormal browsing behavior data of wrong clicks or failure to close the page in time after clicking.

[0014] In the second aspect, the present application provides a book recommendation system based on artificial intelligence, which adopts the following technical solutions:

[0015] A book recommendation system based on artificial intelligence comprises: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned book recommendation method based on artificial intelligence is implemented.

[0016] This application has the following technical effects:

[0017] According to the importance of words in the obtained articles, articles are recommended to users. Based on the full consideration of the reading abilities of different users, a personalized article recommendation method is provided. By analyzing the user's reading behavior, preferences and feedback, the recommendation list is dynamically adjusted so that the longer the user uses the system, the more the recommended articles meet the user's needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present application will become easily understood. In the accompanying drawings, several embodiments of the present application are shown in an exemplary and non-restrictive manner, and the same or corresponding numbers represent the same or corresponding parts.

[0019] Figure 1 This is a method flow chart of an artificial intelligence-based book recommendation method in an embodiment of the present application.

[0020] Figure 2 This is a method flow chart of step S1 in an artificial intelligence-based book recommendation method in an embodiment of the present application.

[0021] Figure 3 This is a method flow chart of step S15 in an artificial intelligence-based book recommendation method in an embodiment of the present application.

[0022] Figure 4 This is a method flow chart of step S4 in an artificial intelligence-based book recommendation method in an embodiment of the present application. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0024] It should be understood that when the terms "first", "second", etc. are used in the claims, specification and drawings of the present application, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the specification and claims of the present application indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their collections.

[0025] The present application embodiment discloses a book recommendation method based on artificial intelligence, referring to Figure 1 , including steps S1 to S5, which are specifically as follows:

[0026] S1: Classify the article and segment the words to obtain the words and their categories, and calculate the importance of the words. Figure 2 , step S1 includes steps S11 to S15, which are specifically as follows:

[0027] S11: Categorize the articles into categories including title, abstract, chapter headings, main text and references.

[0028] Construct a book search system, which includes: a search column and an article library. Taking an article as an example, the article is first divided into categories. The division method can be divided according to the writing structure of the article. In one embodiment, the categories include: title, abstract, chapter title, text and references.

[0029] S12: Segment the text into words according to the word segmentation tool to obtain independent words after segmentation, and determine the category of the words.

[0030] Use word segmentation tools to segment the article, obtain the segmented words, determine the category of the words, and remove unclassified words to eliminate meaningless words.

[0031] In one embodiment, the word segmentation tool may be Jieba word segmentation. Jieba word segmentation is an open source Chinese word segmentation tool that can segment continuous Chinese text into meaningful words required by users. Jieba word segmentation is an existing technology and will not be described in detail here.

[0032] S13: Calculate the first prominence of each word in each category, where the first prominence is the ratio of the number of occurrences of each word to the total number of all words in the category to which each word belongs.

[0033] In one embodiment, taking a word in an article as an example, the ratio of the number of occurrences of the word to the total number of all words in the category to which the word belongs is calculated as the first prominence of the word in the category. According to this method, the first prominence of each word in the category is calculated.

[0034] S14: Calculate the second prominence of each word in the text.

[0035] The calculation formula for the second prominence degree is: ,in, is the second prominence of the word, is the starting point of accumulation, is the accumulation end point, For words in category The first degree of prominence.

[0036] S15: Correct the second prominence level to obtain the word importance level.

[0037] Due to the uneven distribution of words in text, a word may appear densely in certain positions in the article. Clustering the position of a word in the article can obtain the density of a word in certain positions. Figure 3 , step S15 includes steps S151 to S153, which are specifically as follows:

[0038] S151: Sort the words according to the order of the article to obtain the text.

[0039] S152: According to the clustering algorithm, the positions of words in the text are clustered to obtain the maximum cluster of words. One word corresponds to one maximum cluster, wherein a word is set as a target word, the input of clustering is the position of the target word, and the output of clustering is a number of clusters. Each cluster contains multiple target words. The cluster with the largest number of target words is taken as the maximum cluster, and the number of target words is the number of samples.

[0040] The clustering algorithm used in this application is the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm. The clustering radius is the number of two lines of text in the text, and the minimum number of clustering samples is 1. The number of samples represents the position density of a word. The DBSCAN algorithm is a prior art and will not be described in detail here.

[0041] S153: Calculate word importance.

[0042] The calculation formula for word importance is: ,in, is the importance of the word, r is the second prominence of the word, is the number of samples in the largest cluster corresponding to the word, For An exponential function with base .

[0043] In the article, each word corresponds to a word importance. All words are traversed to obtain the word importance of all words, which are recorded as a word importance set. Several articles are obtained in the article library, and each article corresponds to a word importance set.

[0044] In other embodiments, the second prominence degree in step S14 may also be used as the word importance degree. This embodiment is suitable for application scenarios where the distribution of words in the text is relatively uniform.

[0045] S2: Generate a recommendation list based on the importance of the words and obtain the reading time of the users browsing the articles in the recommendation list.

[0046] In one embodiment, a user enters search text in the search bar of a book retrieval system, and repeats the segmentation and elimination steps in step S1 to obtain keywords of the search text. In the word importance set of the article, the words that are the same as the keywords and the word importance corresponding to the words are retained as the recommendation level of the article for the search text.

[0047] Set a degree threshold. When the recommendation degree of an article is greater than the degree threshold, extract the article as a recommendation. The degree threshold of this application can be set to 0.3. Sort the articles in descending order according to the degree of recommendation. Record the user's stay time on each article in the recommendation list as the reading time. The reading time is the time from the user opening the article to closing the article.

[0048] S3: Divide the reading time by the number of words in the article as the user's reading speed.

[0049] Based on the user's browsing behavior, the user's reading speed is calculated. The reading speed is calculated by dividing the reading time by the number of words in the article.

[0050] S4: Calculate the user's interest in the recommendation list. Figure 4 , step S4 includes steps S41 to S44, which are specifically as follows:

[0051] S41: Classify the user's reading speed by time to obtain a reading speed set.

[0052] S42: Using a preset detection model, detect and delete abnormal data in the reading speed set to obtain a reading habit set, wherein the abnormal data includes abnormal browsing behavior data of wrong clicks or failure to close the page in time after clicking.

[0053] Wrong clicks or failure to close the button in time will cause abnormal reading speeds and outliers in the calculation. In order to remove these interfering data, the detection model is used to isolate the abnormal points. The input of the detection model is the reading speed set, and the output is the abnormal reading speed. The abnormal reading speed is deleted from the reading speed set to remove the interfering data.

[0054] The detection model of this application is the LOF (Local Outlier Factor) detection model. The LOF detection model is mainly used to identify abnormal values ​​or outliers in a data set, and measures the degree of abnormality of a data point by calculating the local density of each data point and the difference between adjacent data points. The LOF detection model is a prior art and will not be described in detail here.

[0055] S43: Taking the average of the reading speeds in the reading habit set as the user's reading ability.

[0056] S44: Calculate the interest level.

[0057] The calculation formula for interest level is: ,in, is the user's interest level, For the user's reading speed, For the user's reading ability, For The exponential function with base , represents the minimum function, represents the maximum value function, represents a set of reading habits, Represents the coefficient parameter.

[0058] represents the minimum value in the reading habit set, Indicates the maximum value in the reading habit set. The closer the interest level is to 1, the higher the user's interest in the article. Conversely, the lower the user's interest in the article.

[0059] S5: Obtain word-related evaluations based on the interest level, and in response to a refresh instruction, update the recommendation list based on the word-related evaluations.

[0060] The average of multiple initial word relevance evaluations is taken as the word relevance evaluation.

[0061] The initial word related evaluation is the product of the user's interest in the article he has read and the word's importance. The number of initial word related evaluations corresponding to each word is the same as the number of articles he has read. For example, the word shear force exists in all five articles he has read. Then, the word shear force is set with an initial word related evaluation for each article he has read, that is, there are five initial word related evaluations and the average is calculated to obtain the word related evaluation.

[0062] If the user's refresh command is obtained, the recommendation list is updated according to the word-related evaluation. For example, in the first search, the keyword of the user's search text is architecture, and the recommendation list contains articles related to architecture. Among the articles related to architecture, there are articles related to foundation pits and articles related to shear force, etc. The word-related evaluation of the word shear force and the word-related evaluation of the word foundation pit are calculated, and the word-related evaluation of shear force is greater than the word-related evaluation of foundation pit. After receiving the user's refresh command, the recommendation list is updated so that the ranking of articles related to architecture and foundation pits is lower than the ranking of articles related to architecture and shear force, and the second search result is generated. The second search result is more in line with the user's search needs than the first search result.

[0063] The embodiment of the present application also discloses an artificial intelligence-based book recommendation system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the artificial intelligence-based book recommendation method according to the present application is implemented.

[0064] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0065] In the present application, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module or both. Any such computer storage medium may be part of a device or accessible or connectable to a device.

[0066] Although this specification has shown and described a plurality of embodiments of the present application, it is obvious to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, modifications and alternatives without departing from the thought and spirit of the present application. It should be understood that in the process of practicing the present application, various alternatives to the embodiments of the present application described herein may be adopted.

[0067] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A book recommendation method based on artificial intelligence, characterized in that: Includes steps: Classify the articles and segment the words to obtain the words and their categories, and calculate the importance of the words; the categories include: title, abstract, chapter title, text and references; Generate a recommendation list based on the importance of the words, obtain the reading time of the user browsing the articles in the recommendation list; divide the reading time by the number of words in the article as the user's reading speed; Calculate the user's interest level in the recommendation list; obtain word-related evaluations based on the interest level, and update the recommendation list based on the word-related evaluations in response to a refresh instruction; The method for calculating the interest level is as follows: classify the user's reading speed by time to obtain a reading speed set; use a preset detection model to detect and delete abnormal data in the reading speed set to obtain a reading habit set; use the average of the reading speed in the reading habit set as the user's reading ability; calculate the interest level, and the calculation formula is: ,in, is the user's interest level, For the user's reading speed, For the user's reading ability, For The exponential function with base , represents the minimum function, represents the maximum value function, represents a set of reading habits, represents the coefficient parameter; The product of the user's interest in the article he has read and the word importance is taken as the initial word relevance evaluation, and the average of multiple initial word relevance evaluations is taken as the word relevance evaluation.

2. The artificial intelligence-based book recommendation method according to claim 1, characterized in that: Calculating word importance includes the following steps: Calculate the first prominence of each word in each category. The first prominence is the ratio of the number of occurrences of each word to the total number of all words in the category where each word belongs. Calculate the second prominence of each word in the article, the calculation formula is: ,in, is the second prominence of the word, is the starting point of accumulation, is the accumulation end point, For words in category The first prominence degree; The second prominence level is taken as the word importance level.

3. The book recommendation method based on artificial intelligence according to claim 1, characterized in that: Calculating word importance includes the following steps: Calculate the first prominence of each word in each category. The first prominence is the ratio of the number of occurrences of each word to the total number of all words in the category where each word belongs. Calculate the second prominence of each word in the article, the calculation formula is: ,in, is the second prominence of the word, is the starting point of accumulation, is the accumulation end point, For words in category The first prominence degree; The second prominence level is modified to obtain the word importance level.

4. The artificial intelligence-based book recommendation method according to claim 3, characterized in that: The correction process includes the following steps: According to the order of the article, the words are sorted to obtain the text; According to the clustering algorithm, the positions of words in the text are clustered to obtain the maximum cluster of words. One word corresponds to one maximum cluster. A word is set as the target word. The input of clustering is the position of the target word. The output of clustering is several clusters. Each cluster contains multiple target words. The cluster with the largest number of target words is the largest cluster. The number of target words is the number of samples. The calculation formula for word importance is: ,in, is the importance of the word, The second most prominent degree of the word, is the number of samples in the largest cluster corresponding to the word, For An exponential function with base .

5. The artificial intelligence-based book recommendation method according to any one of claims 1 to 4, characterized in that: Classify the article and segment the words to get the words and the categories they belong to, including the following steps: Categorize the articles into categories including: title, abstract, chapter headings, main text and references; According to the word segmentation tool, the text is segmented into words, independent words are obtained after segmentation, and the category of the word is determined.

6. The artificial intelligence-based book recommendation method according to claim 1, characterized in that: Abnormal data includes: abnormal browsing behavior data such as wrong clicks or failure to close the page in time after clicking.

7. A book recommendation system based on artificial intelligence, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the artificial intelligence-based book recommendation method according to any one of claims 1 to 6 is implemented.

Images