CN117972206A - Content recommendation system, electronic equipment and storage medium based on artificial intelligence - Google Patents
Content recommendation system, electronic equipment and storage medium based on artificial intelligence Download PDFInfo
- Publication number
- CN117972206A CN117972206A CN202410128963.3A CN202410128963A CN117972206A CN 117972206 A CN117972206 A CN 117972206A CN 202410128963 A CN202410128963 A CN 202410128963A CN 117972206 A CN117972206 A CN 117972206A
- Authority
- CN
- China
- Prior art keywords
- user
- data
- recommendation
- unit
- analysis
- 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.)
- Pending
Links
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 23
- 238000004458 analytical method Methods 0.000 claims abstract description 129
- 230000006399 behavior Effects 0.000 claims abstract description 109
- 230000008451 emotion Effects 0.000 claims abstract description 85
- 238000005457 optimization Methods 0.000 claims abstract description 20
- 230000004927 fusion Effects 0.000 claims abstract description 15
- 238000004422 calculation algorithm Methods 0.000 claims description 78
- 238000000034 method Methods 0.000 claims description 40
- 230000010354 integration Effects 0.000 claims description 18
- 230000000694 effects Effects 0.000 claims description 16
- 238000005516 engineering process Methods 0.000 claims description 15
- 230000008859 change Effects 0.000 claims description 14
- 238000007781 pre-processing Methods 0.000 claims description 14
- 238000010801 machine learning Methods 0.000 claims description 12
- 238000013135 deep learning Methods 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 10
- 238000000605 extraction Methods 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 9
- 238000013480 data collection Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 7
- 238000011156 evaluation Methods 0.000 claims description 7
- 230000004913 activation Effects 0.000 claims description 6
- 238000007405 data analysis Methods 0.000 claims description 6
- 238000012544 monitoring process Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 230000004044 response Effects 0.000 claims description 5
- 238000007619 statistical method Methods 0.000 claims description 5
- 238000012731 temporal analysis Methods 0.000 claims description 5
- 238000000700 time series analysis Methods 0.000 claims description 5
- 238000007621 cluster analysis Methods 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 4
- 238000005065 mining Methods 0.000 claims description 4
- 230000002996 emotional effect Effects 0.000 claims description 3
- 238000012549 training Methods 0.000 claims description 3
- 230000003542 behavioural effect Effects 0.000 claims description 2
- 238000004891 communication Methods 0.000 description 5
- 239000013598 vector Substances 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000012300 Sequence Analysis Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000556 factor analysis Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a content recommendation system, electronic equipment and storage medium based on artificial intelligence, which relate to the technical field of content recommendation, wherein the content recommendation system comprises: the system comprises a user behavior analysis module, an emotion analysis module, a context awareness recommendation module, a cross-domain content recommendation module, a dynamic learning adjustment module, a creative content discovery module and a data fusion optimization module. The recommendation accuracy is improved by comprehensively utilizing the user behavior data and emotion analysis, cross-domain recommendation and dynamic learning adjustment are combined to enrich and adapt to diversified demands of users, and meanwhile, the correlation of recommended content and the overall performance of the system are ensured through data fusion and optimization, so that the user satisfaction is remarkably improved.
Description
Technical Field
The present invention relates to the field of content recommendation technologies, and in particular, to a content recommendation system based on artificial intelligence, an electronic device, and a storage medium.
Background
In the digital age, with the popularization of the internet and mobile devices, the channel and mode of information exposure have changed significantly, and particularly in the fields of entertainment, education, information acquisition and the like, personalized content recommendation has become a key technology, and existing content recommendation systems mainly recommend based on the historical behaviors and explicit preferences of users, but these systems often face several limitations and challenges.
Firstly, conventional recommendation systems often cannot accurately capture dynamic changes and diversified demands of users, interests and preferences of users are not static but vary with time, situation and environment, so that relying on historical data and simple algorithms alone often cannot provide satisfactory recommendation effects, secondly, existing recommendation techniques rarely consider the influence of emotional states and cross-domain interests of users, which limits the diversity and depth of recommended contents, and furthermore, due to the lack of effective data integration and analysis mechanisms, recommendation systems often cannot fully utilize a large amount of collected data, resulting in lack of individuation and accuracy of recommendation results.
Aiming at the problems, the invention aims to solve the technical problems of the traditional recommendation system in the aspects of user interest identification, emotion analysis, cross-domain recommendation, dynamic learning and data fusion.
Disclosure of Invention
Aiming at the technical problems in the prior art, the invention provides a content recommendation system, electronic equipment and storage medium based on artificial intelligence, which are used for effectively integrating data from different modules and optimizing the generation process of recommendation results by applying a data analysis technology, thereby improving the correlation of recommendation contents and the performance and user satisfaction of the whole system.
According to a first aspect of the present invention, there is provided an artificial intelligence based content recommendation system comprising:
user behavior analysis module: for collecting online behavior data of a user, analyzing the behavioral data to understand interests and preferences of the user;
Emotion analysis module: based on the analysis data of the user behavior analysis module, analyzing the emotion tendency of the user in social media or comments, and providing additional user interest dimension for a recommendation system;
Context aware recommendation module: receiving data from a user behavior analysis module and an emotion analysis module, and providing more accurate content recommendation for a user based on the current environment and situation of the user;
A cross-domain content recommendation module: analyzing behavior data of users in different fields to find potential interest intersections and recommending cross-domain content;
dynamic learning adjustment module: tracking the recommendation effect in real time, and continuously adjusting a recommendation algorithm according to feedback and behavior change of a user;
Creative content mining module: analyzing minority content and emerging trends using machine learning algorithms to recommend non-mainstream but attractive content to a user;
And a data fusion and optimization module: and integrating and analyzing the data collected by the modules so as to optimize the accuracy and the relativity of the recommended result.
On the basis of the technical scheme, the invention can also make the following improvements.
Optionally, the user behavior analysis module includes:
A data collection unit: the method comprises the steps of acquiring online behavior data of a user, wherein the online behavior data comprise user browsing history, click rate and residence time;
A data preprocessing unit: the data preprocessing module is used for preprocessing the collected behavior data;
feature extraction unit: for extracting key features from the preprocessed data;
Deep learning analysis unit: for analyzing the extracted features;
User portrait construction unit: constructing a user portrait based on the output of the deep learning analysis unit, and providing a basis for recommending contents; the formula for specifically constructing the user portrait is as follows: p=u×w p+bp, where P denotes a user image, U is a user interest analysis result obtained from the deep learning analysis unit, and W p and b p are weights and deviations for user image construction.
Optionally, the emotion analysis module includes:
Emotion data association unit: data for receiving user interests and preferences from the user behavior analysis module, including topics that are frequently accessed and interacted with by the user, keywords, and for guiding the collection of social media or comment text related to the user interests;
text preprocessing unit: carrying out standardization processing on the collected text data, and preparing formatted text for subsequent emotion analysis;
emotion feature extraction section: extracting emotion related features from the preprocessed text, wherein the features comprise keyword frequency of user interest, appearance frequency of emotion vocabulary and emotion polarity of sentences;
emotion analysis algorithm unit: analyzing the extracted features to determine emotional tendency of the text;
A comprehensive emotion scoring unit: and calculating the comprehensive emotion score of the user according to the emotion analysis result so as to reflect the emotion tendency of the user in the specific interest area.
Optionally, the context-aware recommendation module includes:
a data integration unit: the system comprises a user behavior analysis module, a user emotion trend analysis module and a user emotion trend analysis module, wherein the user behavior analysis module is used for receiving and merging user behavior data from the user behavior analysis module, including browsing history, click rate and user emotion trend data of the emotion analysis module, and integrating the collected data to ensure that data from different sources can be seamlessly combined to form a comprehensive user portrait;
An environment sensing unit: the method comprises the steps of collecting relevant data of the current environment of a user, including time data, place data and equipment type data;
a context analysis unit: deep analysis is carried out on the data collected by the environment sensing unit, and a cluster analysis algorithm is adopted to identify the specific situation of the user;
Recommendation algorithm unit: and generating personalized content recommendation for the user by using a recommendation algorithm in combination with the output of the comprehensive user portrait and the context analysis unit provided by the data integration unit, wherein the recommendation algorithm formula is as follows:
Wherein R represents a recommendation result, f is an activation function for processing the output of the recommendation, W r and b r are the weight and deviation of the recommendation model, respectively, and X b、Xs and X c represent user behavior data, emotion data and context data,/>, respectively Representing the integration and fusion of data.
Optionally, the cross-domain content recommendation module includes:
Behavior data collection unit: the method is used for collecting behavior data of the user in different fields, and is specifically classified as follows:
for the video domain, the collected data includes viewing history, viewing duration, and user feedback;
For the music domain, the collected data includes song listening frequency, playlist, and favorite song types;
For the reading field, the collected data comprise the read books, the chapter completion rate and the reading time;
behavior pattern analysis unit: analyzing user behavior patterns in different fields by using a statistical analysis method, wherein the statistical analysis method comprises the steps of identifying interest points of users by calculating and comparing similarity and frequency of user activities in each field;
an interest cross finding unit: identifying intersections of user behavior patterns in different domains using an applied association rule learning algorithm based on the results of the behavior pattern analysis unit to find potential points of interest;
a cross-domain recommendation algorithm unit: based on the found interest intersection points and the behavior modes of the user in each field, a preset recommendation algorithm is adopted to generate cross-domain content recommendation.
Optionally, the dynamic learning adjustment module includes:
a recommended effect tracking unit: the method comprises the steps of monitoring the response of a user to recommended content, wherein the response comprises the click rate, the watching or listening time length and the page browsing depth of the user to the recommended content;
User feedback analysis unit: the method comprises the steps of analyzing feedback information directly provided by a user, including scoring and commenting on recommended content, and extracting emotion tendency and preference information from text feedback of the user by applying a text analysis technology;
Behavior change detection unit: for monitoring the trend of changes in user behavior, including the appearance of new points of interest or a decrease in the degree of interest for certain content types, identifying changes in the user behavior pattern using a time series analysis method, and indicating the evolution of user interests or preferences;
A recommendation algorithm adjusting unit: in combination with the data provided by the above units, a machine learning algorithm is used to continuously adjust and optimize the recommendation algorithm, wherein the formula of the adjustment algorithm is as follows:
Wherein: w new is the updated weight, W old is the original weight, α is the learning rate,/> Is the gradient of the loss function with respect to the weight, and X and Y represent the input data and the user's feedback, respectively.
Optionally, the creative content discovery module includes:
A minority content identification unit: the method is used for identifying and classifying the content which is not widely contacted by the user, and adopts a machine learning algorithm, particularly a clustering algorithm to identify non-mainstream trends in the user behavior data;
emerging trend analysis unit: identifying emerging content trends using time series analysis and prediction models, including assessing changes in growth rate and user engagement of content;
creative content recommending unit: and generating the recommendation of the creative content for the user by using a recommendation algorithm according to the analysis results of the minority content identification unit and the emerging trend analysis unit.
Optionally, the data fusion and optimization module includes:
A data integration unit: for collecting data from various modules within the system, including user behavior data,
Emotion analysis data, context data, cross-domain content data and dynamic learning adjustment data, and uniformly formatting the collected data to create a comprehensive data set;
Multidimensional analysis unit: based on data integration, a multidimensional data analysis technology is applied to identify and extract key characteristics of user behaviors and preferences;
Recommendation algorithm optimizing unit: optimizing a recommendation model by using a machine learning algorithm in combination with the multidimensional analysis result, wherein the optimization process comprises the steps of adjusting model parameters, training a model and evaluating model performance;
Recommendation result evaluation unit: and evaluating the output of the recommendation algorithm after optimization, and measuring the quality of the recommendation result by using the evaluation indexes of the accuracy rate, the recall rate and the F1 score.
The invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements an artificial intelligence based content recommendation system as described in any one of the above when executing the program.
The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements an artificial intelligence based content recommendation system as described in any of the above.
The invention has the technical effects and advantages that:
According to the content recommendation system, the electronic equipment and the storage medium based on the artificial intelligence, the real-time requirements and the preferences of the user can be more accurately understood through comprehensively analyzing the behavior data and the emotion tendencies of the user, and the depth analysis enables the recommended content to be closer to the current interests of the user, so that the accuracy of the recommendation system is greatly improved.
Meanwhile, through analyzing the behavior patterns of the user in different fields, the system can find potential cross-domain interest points of the user, so that the experience of the user is enriched, and more diversified content selection is provided for the user. The recommendation algorithm can be adjusted in real time according to the change of the user behavior and the feedback information by utilizing the dynamic learning adjustment module, and the continuous learning and optimization process ensures that the recommendation system can flexibly adapt to the evolution of the user interest and continuously provide high-quality recommendation content. By effectively integrating the data from different modules and applying the data analysis technology, the generation process of the recommendation result is optimized, so that the correlation of the recommendation content is improved, and the performance and the user satisfaction of the whole system are improved.
Drawings
Fig. 1 is a schematic diagram of a content recommendation system according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
It can be appreciated that based on the defects in the background technology, the embodiment of the invention provides a content recommendation system based on artificial intelligence, as shown in fig. 1, which comprises a user behavior analysis module, an emotion analysis module, a context perception recommendation module, a cross-domain content recommendation module, a dynamic learning adjustment module, a creative content mining module and a data fusion optimization module; wherein,
User behavior analysis module: the online behavior data of the user are collected, and the interests and the preferences of the user are understood through analyzing the behavior data through a deep learning algorithm;
Emotion analysis module: based on the analysis data of the user behavior analysis module, analyzing the emotion tendency of the user in social media or comments by using a natural language processing technology, and providing additional user interest dimension for a recommendation system;
context aware recommendation module: receiving data from the user behavior analysis module and the emotion analysis module, and providing more accurate content recommendation for the user based on the current environment and situation of the user, such as time, place and device type;
A cross-domain content recommendation module: analyzing behavior data of users in different fields (such as video, music, reading and the like) to find potential interest intersections and recommending cross-domain content;
dynamic learning adjustment module: tracking the recommendation effect in real time, and continuously adjusting a recommendation algorithm according to feedback and behavior change of a user;
Creative content mining module: analyzing minority content and emerging trends using machine learning algorithms to recommend non-mainstream but attractive content to a user;
And a data fusion and optimization module: and integrating and analyzing the data collected by the modules so as to optimize the accuracy and the relativity of the recommended result.
Further, the user behavior analysis module comprises a data collection unit, a data preprocessing unit, a feature extraction unit, a deep learning analysis unit and a user portrait construction unit; wherein,
A data collection unit: directly acquiring online behavior data of a user through a webpage tracking technology and an application program interface, wherein the behavior data comprises user browsing history, click rate and residence time;
A data preprocessing unit: the collected behavior data is subjected to data preprocessing so as to facilitate further analysis, wherein the preprocessing comprises the steps of data denoising, missing value processing and normalization processing, so that the quality and consistency of the data are ensured;
Feature extraction unit: extracting key features from the preprocessed data, wherein the key features comprise page access frequency F, the click times C of a user on specific content and average residence time T;
deep learning analysis unit: the extracted features are analyzed using Convolutional Neural Networks (CNNs) to understand the interests and preferences of the user, specifically using the following formulas: u=f (W x+b), where U represents the user interest analysis result, f is an activation function, such as ReLU or Sigmoid, W and b are the weights and biases of the convolutional neural network, respectively, and X is the input eigenvector, including F, C, T;
User portrait construction unit: constructing a user portrait based on the output of the deep learning analysis unit, wherein the portrait comprises information of interest fields, preference types and potential requirements of users, and provides a basis for recommending contents, and a formula for specifically constructing the user portrait is as follows: p=u×w p+bp, where P denotes a user image, U is a user interest analysis result obtained from the deep learning analysis unit, and W p and b p are weights and deviations for user image construction.
Further, the emotion analysis module comprises an emotion data association unit, a text preprocessing unit, an emotion feature extraction unit, an emotion analysis algorithm unit and a comprehensive emotion scoring unit; wherein,
Emotion data association unit: data for receiving user interests and preferences from the user behavior analysis module, including topics, keywords that are frequently accessed and interacted with by the user, and for guiding the collection of social media or comment text related to the user interests;
text preprocessing unit: the method comprises the steps of carrying out standardization processing on collected text data, including removal of irrelevant characters, text segmentation and part-of-speech tagging, and preparing formatted text for subsequent emotion analysis;
emotion feature extraction section: extracting emotion related features from the preprocessed text, wherein the features comprise keyword frequencies of user interests, occurrence frequencies of emotion vocabularies and emotion polarities of sentences;
Emotion analysis algorithm unit: analyzing the extracted features by using natural language processing technology to determine the emotion tendencies of the text, wherein the emotion analysis formula is as follows: s=f (W e·Xe+be), where S represents the emotion analysis result, f is the activation function of emotion analysis, W e and b e are the weight and deviation of the emotion analysis model, respectively, and X e is the input feature vector, based on the features extracted by the emotion feature extraction unit;
a comprehensive emotion scoring unit: according to the emotion analysis result, calculating the comprehensive emotion score of the user so as to reflect the emotion tendency of the user in the specific interest area, wherein the emotion score has the following calculation formula:
Where E is the composite emotion score, S i is the emotion analysis result for a single text, and w i is the weight associated with each S i.
Further, the context-aware recommendation module comprises a data integration unit, an environment awareness unit, a context analysis unit and a recommendation algorithm unit; wherein,
A data integration unit: the system comprises a user behavior analysis module, a user behavior data acquisition module, a user behavior analysis module and a user emotion trend analysis module, wherein the user behavior data comprises browsing history, click rate and user emotion trend data of the emotion analysis module;
An environment sensing unit: relevant data for collecting the current environment of the user, including time data (determining the time period of the user's activity), location data (acquiring the user's precise location using GPS or network information), and device type data (identifying the type of device the user is using, such as a smart phone, tablet computer, or notebook computer);
A context analysis unit: deep analysis is carried out on the data collected by the environment sensing unit, a cluster analysis algorithm is specifically adopted to identify the specific situation of the user, for example, time and place data are analyzed to infer that the user is likely to be in a working state or a leisure state, and equipment type data are used for inferring the use habit and preference of the user;
recommendation algorithm unit: and generating personalized content recommendation for the user by using a recommendation algorithm by combining the comprehensive user portrait and the output of the situation analysis unit provided by the data integration unit, wherein the recommendation algorithm has the following formula:
Wherein R represents a recommendation result, f is an activation function for processing the output of the recommendation, W r and b r are the weight and deviation of the recommendation model, respectively, and X b、Xs and X c represent user behavior data, emotion data and context data,/>, respectively Representing the integration and fusion of data.
In the above technical solution, the specific steps for identifying the specific situation where the user is located by using the cluster analysis algorithm in the situation analysis unit are as follows:
S1: firstly, receiving data from an environment sensing unit, setting time data as T, place data as L and equipment type data as D, and forming the data into feature vectors X c = [ T, L, D ];
S2: processing the feature vectors to adapt to a clustering algorithm, specifically normalizing the time and place data for comparison on the same scale;
S3: identifying the situation of the user by using a Kmeans clustering algorithm; the formula of the clustering algorithm is specifically as follows:
Where C i is the cluster center assigned to a particular user context, μ j is an element in the cluster center set M, X is an element in the feature vector X c, and X- μ j∥2 is the square of the Euclidean distance used to calculate the distance between the feature vector X and the cluster center μ j;
S4: each cluster center represents a specific user context, and the algorithm in S3 will determine the context in which the user is located by analyzing the distance between the user' S environmental data and the respective cluster center.
Further, the cross-domain content recommendation module comprises a behavior data collection unit, a behavior pattern analysis unit, an interest cross discovery unit and a cross-domain recommendation algorithm unit; wherein,
Behavior data collection unit: the method is used for collecting behavior data of the user in different fields, and is specifically classified as follows:
for the video domain, the collected data includes viewing history, viewing duration, and user feedback;
For the music domain, the collected data includes song listening frequency, playlist, and favorite song types;
For the reading field, the collected data comprise the read books, the chapter completion rate and the reading time;
behavior pattern analysis unit: the statistical analysis method is used for analyzing the user behavior patterns in different fields, and the interest points of the user are identified by calculating and comparing the similarity and the frequency of the user activities in each field, wherein the specific formula is as follows: Wherein M represents the average behavior mode of the user in a specific field, n is the activity number of the user in the field, X i is the behavior data (such as watching time length and playing times) of single activity, and the calculation formula aims at calculating the average activity level of the user in each field and provides a basis for identifying interest points;
An interest cross finding unit: based on the results of the behavior pattern analysis unit, an intersection of user behavior patterns in different domains is identified using an applied association rule learning algorithm to find potential points of interest, for example, if a user frequently views a history record in the video domain and prefers a history book in the reading domain, the system identifies the user's cross-interests in the history content, and the specific identification formula is as follows: wherein C represents an interest intersection, M k is a behavior pattern of the kth domain, and K is the total number of domains in which the user is active;
a cross-domain recommendation algorithm unit: based on the found interest intersection points and the behavior patterns of the user in each field, a preset recommendation algorithm is adopted to generate cross-domain content recommendation, and the recommendation algorithm has the following formula:
Where R is the final recommendation, f r is the activation function for processing the output of the recommendation, W r and b r are the weights and deviations of the recommendation model, respectively, M and C represent the behavior pattern and the interest intersection of the user, respectively,/> Representing a fusion of the data.
Further, the dynamic learning adjustment module comprises a recommendation effect tracking unit, a user feedback analysis unit, a behavior change detection unit and a recommendation algorithm adjustment unit; wherein,
A recommended effect tracking unit: the method is used for monitoring the response of the user to the recommended content and specifically comprises the click rate, the watching or listening time length and the page browsing depth of the recommended content, wherein the click rate is high, the watching time length and the depth browsing indicate the interest and satisfaction degree of the user to the recommended content;
User feedback analysis unit: the method comprises the steps of analyzing feedback information directly provided by a user, including scoring and commenting on recommended content, extracting emotion tendency and preference information from text feedback of the user by applying text analysis technology, and providing quantitative indexes for solving satisfaction degree of the user on the recommended content;
Behavior change detection unit: the method comprises the steps of monitoring the change trend of user behaviors, including the occurrence of new interest points or the decrease of attention to certain content types, identifying the change in the user behavior mode by using a time sequence analysis method and indicating the evolution of user interests or preferences, wherein the change provides a data basis for the adjustment of a recommendation system;
A recommendation algorithm adjusting unit: in combination with the data provided by the above units, the recommendation algorithm adjusting unit uses a machine learning algorithm to continuously adjust and optimize the recommendation algorithm, and the formula of the adjustment algorithm is as follows:
Wherein: w new is the updated weight, W old is the original weight, α is the learning rate,/> Is the gradient of the loss function with respect to the weight, and X and Y represent the input data and the user's feedback, respectively.
Further, the creative content discovery module comprises a minority content identification unit, an emerging trend analysis unit and a creative content recommendation unit; wherein,
A minority content identification unit: the method is used for identifying and classifying the content which is not widely contacted by the user, and particularly adopts a machine learning algorithm, particularly a clustering algorithm to identify the non-mainstream trend in the user behavior data, wherein the clustering algorithm has the following formula: Wherein, C min represents the identified minority content category, n is the total number of data points, x i is the number of data points representing the user behavior, x i is a single user behavior data point, including the interactive data of the user on the specific content, C j is the clustering center, represents the center point of the specific content category, β is an adjustment parameter for controlling the degree of importance of the separation between the categories in the clustering process, and improving the identification accuracy of the minority content;
Emerging trend analysis unit: the time series analysis and prediction model is used to identify emerging content trends, and the process comprises evaluating the growth rate of the content and the change of the participation of the user, wherein the specific calculation formula is as follows:
T new＝γ·(Xcurrent-Xprev)+Tprev, where T new is a new trend score, γ is a change sensitivity coefficient for adjusting the sensitivity of the new trend identification to changes in user behavior, X current and X prev represent user engagement data at current and previous time points, such as click, view, comment, etc., and T prev is a trend score at the previous time point;
Creative content recommending unit: and generating the recommendation of the creative content for the user by using a recommendation algorithm in combination with the analysis results of the minority content identification unit and the emerging trend analysis unit, wherein the recommendation algorithm combines the scores of the minority content and the emerging trend to optimize the diversity and the attraction of the recommendation.
Further, the data fusion and optimization module comprises a data integration unit, a multidimensional analysis unit, a recommendation algorithm optimization unit and a recommendation result evaluation unit; wherein,
A data integration unit: for collecting data from various modules within the system, including user behavior data, emotion analysis data, context data, cross-domain content data, and dynamic learning adjustment data, the collected data are formatted uniformly, a comprehensive data set is created, and a basis is provided for subsequent analysis and optimization;
Multidimensional analysis unit: on the basis of data integration, multidimensional data analysis technologies such as Principal Component Analysis (PCA) or factor analysis are applied to identify and extract key features of user behaviors and preferences, which are helpful for reducing the dimensionality of data, while retaining the most influential user features;
recommendation algorithm optimizing unit: optimizing the recommended model by using a machine learning algorithm, such as a random forest or gradient elevator, in combination with the multidimensional analysis result, wherein the optimization process comprises the steps of adjusting model parameters, training the model and evaluating the performance of the model;
Recommendation result evaluation unit: and evaluating the output of the recommendation algorithm after optimization, and measuring the quality of a recommendation result by using the evaluation indexes of the accuracy, the recall rate and the F1 score, so as to ensure that the recommendation content is relevant and accurately meets the requirements of users.
In summary, the embodiment of the invention provides a content recommendation system based on artificial intelligence, which can more accurately understand the real-time requirements and preferences of users by comprehensively analyzing the behavior data and emotion tendencies of the users, and the depth analysis enables the recommended content to be closer to the current interests of the users, so that the accuracy of the recommendation system is greatly improved. Meanwhile, through analyzing the behavior patterns of the user in different fields, the system can find potential cross-domain interest points of the user, so that the experience of the user is enriched, and more diversified content selection is provided for the user. The recommendation algorithm can be adjusted in real time according to the change of the user behavior and the feedback information by utilizing the dynamic learning adjustment module, and the continuous learning and optimization process ensures that the recommendation system can flexibly adapt to the evolution of the user interest and continuously provide high-quality recommendation content. By effectively integrating the data from different modules and applying the data analysis technology, the generation process of the recommendation result is optimized, so that the correlation of the recommendation content is improved, and the performance and the user satisfaction of the whole system are improved.
According to a second aspect of the present invention, there is provided an electronic device, which may include: a processor (processor), a communication interface (Communications Interface), a memory (memory), and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus. The processor may invoke logic instructions in the memory to perform the content of one of the artificial intelligence based content recommendation systems described above.
Further, the logic instructions in the memory described above may be implemented in the form of software functional units and stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In yet another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, is implemented to perform the content of an artificial intelligence based content recommendation system as described above.
In summary, the recommendation accuracy is improved by comprehensively utilizing the user behavior data and emotion analysis, cross-domain recommendation and dynamic learning adjustment are combined to enrich and adapt to diversified demands of users, and meanwhile, the correlation of recommended content and the overall performance of the system are ensured through data fusion and optimization, so that the user satisfaction is remarkably improved.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. An artificial intelligence based content recommendation system, comprising:
user behavior analysis module: for collecting online behavior data of a user, analyzing the behavioral data to understand interests and preferences of the user;
Emotion analysis module: based on the analysis data of the user behavior analysis module, analyzing the emotion tendency of the user in social media or comments, and providing additional user interest dimension for a recommendation system;
Context aware recommendation module: receiving data from a user behavior analysis module and an emotion analysis module, and providing more accurate content recommendation for a user based on the current environment and situation of the user;
A cross-domain content recommendation module: analyzing behavior data of users in different fields to find potential interest intersections and recommending cross-domain content;
dynamic learning adjustment module: tracking the recommendation effect in real time, and continuously adjusting a recommendation algorithm according to feedback and behavior change of a user;
Creative content mining module: analyzing minority content and emerging trends using machine learning algorithms to recommend non-mainstream but attractive content to a user;
And a data fusion and optimization module: and integrating and analyzing the data collected by the modules so as to optimize the accuracy and the relativity of the recommended result.
2. The artificial intelligence based content recommendation system of claim 1 wherein said user behavior analysis module comprises:
A data collection unit: the method comprises the steps of acquiring online behavior data of a user, wherein the online behavior data comprise user browsing history, click rate and residence time;
A data preprocessing unit: the data preprocessing module is used for preprocessing the collected behavior data;
feature extraction unit: for extracting key features from the preprocessed data;
Deep learning analysis unit: for analyzing the extracted features;
User portrait construction unit: constructing a user portrait based on the output of the deep learning analysis unit, and providing a basis for recommending contents; the formula for specifically constructing the user portrait is as follows: p=u×w p+bp, where P denotes a user image, U is a user interest analysis result obtained from the deep learning analysis unit, and W p and b p are weights and deviations for user image construction.
3. The artificial intelligence based content recommendation system of claim 1 wherein said emotion analysis module comprises:
Emotion data association unit: data for receiving user interests and preferences from the user behavior analysis module, including topics that are frequently accessed and interacted with by the user, keywords, and for guiding the collection of social media or comment text related to the user interests;
text preprocessing unit: carrying out standardization processing on the collected text data, and preparing formatted text for subsequent emotion analysis;
emotion feature extraction section: extracting emotion related features from the preprocessed text, wherein the features comprise keyword frequency of user interest, appearance frequency of emotion vocabulary and emotion polarity of sentences;
emotion analysis algorithm unit: analyzing the extracted features to determine emotional tendency of the text;
A comprehensive emotion scoring unit: and calculating the comprehensive emotion score of the user according to the emotion analysis result so as to reflect the emotion tendency of the user in the specific interest area.
4. The artificial intelligence based content recommendation system of claim 1 wherein said context aware recommendation module comprises:
a data integration unit: the system comprises a user behavior analysis module, a user emotion trend analysis module and a user emotion trend analysis module, wherein the user behavior analysis module is used for receiving and merging user behavior data from the user behavior analysis module, including browsing history, click rate and user emotion trend data of the emotion analysis module, and integrating the collected data to ensure that data from different sources can be seamlessly combined to form a comprehensive user portrait;
An environment sensing unit: the method comprises the steps of collecting relevant data of the current environment of a user, including time data, place data and equipment type data;
a context analysis unit: deep analysis is carried out on the data collected by the environment sensing unit, and a cluster analysis algorithm is adopted to identify the specific situation of the user;
Recommendation algorithm unit: and generating personalized content recommendation for the user by using a recommendation algorithm in combination with the output of the comprehensive user portrait and the context analysis unit provided by the data integration unit, wherein the recommendation algorithm formula is as follows:
Wherein R represents a recommendation result, f is an activation function for processing the output of the recommendation, W r and b r are the weight and deviation of the recommendation model, respectively, and X b、Xs and X c represent user behavior data, emotion data and context data,/>, respectively Representing the integration and fusion of data.
5. The artificial intelligence based content recommendation system of claim 1 wherein the cross-domain content recommendation module comprises:
Behavior data collection unit: the method is used for collecting behavior data of the user in different fields, and is specifically classified as follows:
for the video domain, the collected data includes viewing history, viewing duration, and user feedback;
For the music domain, the collected data includes song listening frequency, playlist, and favorite song types;
For the reading field, the collected data comprise the read books, the chapter completion rate and the reading time;
behavior pattern analysis unit: analyzing user behavior patterns in different fields by using a statistical analysis method, wherein the statistical analysis method comprises the steps of identifying interest points of users by calculating and comparing similarity and frequency of user activities in each field;
an interest cross finding unit: identifying intersections of user behavior patterns in different domains using an applied association rule learning algorithm based on the results of the behavior pattern analysis unit to find potential points of interest;
a cross-domain recommendation algorithm unit: based on the found interest intersection points and the behavior modes of the user in each field, a preset recommendation algorithm is adopted to generate cross-domain content recommendation.
6. The artificial intelligence based content recommendation system of claim 1 wherein the dynamic learning adjustment module comprises:
a recommended effect tracking unit: the method comprises the steps of monitoring the response of a user to recommended content, wherein the response comprises the click rate, the watching or listening time length and the page browsing depth of the user to the recommended content;
User feedback analysis unit: the method comprises the steps of analyzing feedback information directly provided by a user, including scoring and commenting on recommended content, and extracting emotion tendency and preference information from text feedback of the user by applying a text analysis technology;
Behavior change detection unit: for monitoring the trend of changes in user behavior, including the appearance of new points of interest or a decrease in the degree of interest for certain content types, identifying changes in the user behavior pattern using a time series analysis method, and indicating the evolution of user interests or preferences;
A recommendation algorithm adjusting unit: in combination with the data provided by the above units, a machine learning algorithm is used to continuously adjust and optimize the recommendation algorithm, wherein the formula of the adjustment algorithm is as follows:
Wherein: w new is the updated weight, W old is the original weight, α is the learning rate,/> Is the gradient of the loss function with respect to the weight, and X and Y represent the input data and the user's feedback, respectively.
7. The artificial intelligence based content recommendation system of claim 1 wherein the creative content discovery module comprises:
A minority content identification unit: the method is used for identifying and classifying the content which is not widely contacted by the user, and adopts a machine learning algorithm, particularly a clustering algorithm to identify non-mainstream trends in the user behavior data;
emerging trend analysis unit: identifying emerging content trends using time series analysis and prediction models, including assessing changes in growth rate and user engagement of content;
creative content recommending unit: and generating the recommendation of the creative content for the user by using a recommendation algorithm according to the analysis results of the minority content identification unit and the emerging trend analysis unit.
8. The artificial intelligence based content recommendation system of claim 1 wherein the data fusion and optimization module comprises:
A data integration unit: the system comprises a system and a method for collecting data from various modules in the system, wherein the data comprises user behavior data, emotion analysis data, context data, cross-domain content data and dynamic learning adjustment data, and the collected data is uniformly formatted to create a comprehensive data set;
Multidimensional analysis unit: based on data integration, a multidimensional data analysis technology is applied to identify and extract key characteristics of user behaviors and preferences;
Recommendation algorithm optimizing unit: optimizing a recommendation model by using a machine learning algorithm in combination with the multidimensional analysis result, wherein the optimization process comprises the steps of adjusting model parameters, training a model and evaluating model performance;
Recommendation result evaluation unit: and evaluating the output of the recommendation algorithm after optimization, and measuring the quality of the recommendation result by using the evaluation indexes of the accuracy rate, the recall rate and the F1 score.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements an artificial intelligence based content recommendation system according to any one of claims 1 to 8 when the program is executed by the processor.
10. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor implements an artificial intelligence based content recommendation system according to any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410128963.3A CN117972206A (en) | 2024-01-30 | 2024-01-30 | Content recommendation system, electronic equipment and storage medium based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410128963.3A CN117972206A (en) | 2024-01-30 | 2024-01-30 | Content recommendation system, electronic equipment and storage medium based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117972206A true CN117972206A (en) | 2024-05-03 |
Family
ID=90862373
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410128963.3A Pending CN117972206A (en) | 2024-01-30 | 2024-01-30 | Content recommendation system, electronic equipment and storage medium based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117972206A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118194950A (en) * | 2024-05-14 | 2024-06-14 | 吉林省科技创新平台管理中心 | Intelligent interaction method and system based on natural language processing |
-
2024
- 2024-01-30 CN CN202410128963.3A patent/CN117972206A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118194950A (en) * | 2024-05-14 | 2024-06-14 | 吉林省科技创新平台管理中心 | Intelligent interaction method and system based on natural language processing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111444428B (en) | Information recommendation method and device based on artificial intelligence, electronic equipment and storage medium | |
| WO2020228514A1 (en) | Content recommendation method and apparatus, and device and storage medium | |
| CN108829822B (en) | Media content recommendation method and device, storage medium and electronic device | |
| CN110781321B (en) | Multimedia content recommendation method and device | |
| CN107992531A (en) | News personalization intelligent recommendation method and system based on deep learning | |
| CA2959835A1 (en) | Sentiment rating system and method | |
| CN113378565B (en) | Event analysis method, device and equipment for multi-source data fusion and storage medium | |
| CN112131472A (en) | Information recommendation method and device, electronic equipment and storage medium | |
| CN117972206A (en) | Content recommendation system, electronic equipment and storage medium based on artificial intelligence | |
| CN113574522A (en) | Selective presentation of rich experiences in a search | |
| Wang et al. | A learning resource recommendation algorithm based on online learning sequential behavior | |
| CN112989179B (en) | Model training and multimedia content recommendation method and device | |
| CN107103093B (en) | Short text recommendation method and device based on user behavior and emotion analysis | |
| Patel et al. | A comparative study of music recommendation systems | |
| CN110727871A (en) | Multi-mode data acquisition and comprehensive analysis platform based on convolution decomposition depth model | |
| Ashraf et al. | Personalized news recommendation based on multi-agent framework using social media preferences | |
| CN113407772B (en) | Video recommendation model generation method, video recommendation method and device | |
| CN114817692A (en) | Method, device and equipment for determining recommended object and computer storage medium | |
| Procter et al. | Enabling social media research through citizen social science | |
| CN116541486A (en) | News information aggregation method based on data mining and deep learning | |
| CN116956183A (en) | Multimedia resource recommendation method, model training method, device and storage medium | |
| US20210390150A1 (en) | Methods and systems for self-tuning personalization engines in near real-time | |
| Babu et al. | Multiclass Text Emotion Recognition in Social Media Data | |
| Zhang | Design of the piano score recommendation image analysis system based on the big data and convolutional neural network | |
| CN116523024B (en) | Training method, device, equipment and storage medium of recall 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 |