Pub Date : 2023-12-01DOI: 10.53106/199115992023123406007
Cui-Cui Cai Cui-Cui Cai, Mao-Sheng Fu Cui-Cui Cai, Xian-Meng Meng Mao-Sheng Fu, Qi-Jian Wang Xian-Meng Meng, Yue-Qin Wang Qi-Jian Wang
As a novel metaheuristic algorithm, the Harris Hawks Optimization (HHO) algorithm has excellent search capability. Similar to other metaheuristic algorithms, the HHO algorithm has low convergence accuracy and easily traps in local optimal when dealing with complex optimization problems. A modified Harris Hawks optimization (MHHO) algorithm with multiple strategies is presented to overcome this defect. First, chaotic mapping is used for population initialization to select an appropriate initiation position. Then, a novel nonlinear escape energy update strategy is presented to control the transformation of the algorithm phase. Finally, a nonlinear control strategy is implemented to further improve the algorithm’s efficiency. The experimental results on benchmark functions indicate that the performance of the MHHO algorithm outperforms other algorithms. In addition, to validate the performance of the MHHO algorithm in solving engineering problems, the proposed algorithm is applied to an indoor visible light positioning system, and the results show that the high precision positioning of the MHHO algorithm is obtained.
{"title":"Modified Harris Hawks Optimization Algorithm with Multi-strategy for Global Optimization Problem","authors":"Cui-Cui Cai Cui-Cui Cai, Mao-Sheng Fu Cui-Cui Cai, Xian-Meng Meng Mao-Sheng Fu, Qi-Jian Wang Xian-Meng Meng, Yue-Qin Wang Qi-Jian Wang","doi":"10.53106/199115992023123406007","DOIUrl":"https://doi.org/10.53106/199115992023123406007","url":null,"abstract":"As a novel metaheuristic algorithm, the Harris Hawks Optimization (HHO) algorithm has excellent search capability. Similar to other metaheuristic algorithms, the HHO algorithm has low convergence accuracy and easily traps in local optimal when dealing with complex optimization problems. A modified Harris Hawks optimization (MHHO) algorithm with multiple strategies is presented to overcome this defect. First, chaotic mapping is used for population initialization to select an appropriate initiation position. Then, a novel nonlinear escape energy update strategy is presented to control the transformation of the algorithm phase. Finally, a nonlinear control strategy is implemented to further improve the algorithm’s efficiency. The experimental results on benchmark functions indicate that the performance of the MHHO algorithm outperforms other algorithms. In addition, to validate the performance of the MHHO algorithm in solving engineering problems, the proposed algorithm is applied to an indoor visible light positioning system, and the results show that the high precision positioning of the MHHO algorithm is obtained.","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"24 12","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139188910","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In recent years, the rapid development of service-oriented computing technology has increased the burden of choice for software developers when developing service-based applications. Existing Web service recommendation systems often face two challenges. First, developers are required to input keywords for service search, but due to their lack of knowledge in the relevant field, the keywords entered by the developers are usually freestyle, causing an inability to accurately locate services. Second, it is exceedingly difficult to extract services that meet the requirements due to the 99.8% sparseness of the application service interaction records. To address the above challenges, a framework for service recommendation through multi-model fusion (SRM) is proposed in this paper. Firstly, we employ graph neural network algorithms to deeply mine historical records, extract the features of applications and services, and calculate their preferences. Secondly, we use the BERT model to analyze text information and use the attention mechanism and fully connected neural networks to deeply mine the matching degree between candidate services and development requirements. The two models mentioned above are further merged to obtain the final service recommendation list. Extensive experiments on datasets demonstrate that SRM can significantly enhance the effectiveness of recommendations in service recommendation scenarios.
{"title":"Service Recommendation Method based on Multi Model Fusion","authors":"Ting Yu Ting Yu, Lihua Zhang Ting Yu, Hongbing Liu Lihua Zhang","doi":"10.53106/199115992023123406005","DOIUrl":"https://doi.org/10.53106/199115992023123406005","url":null,"abstract":"In recent years, the rapid development of service-oriented computing technology has increased the burden of choice for software developers when developing service-based applications. Existing Web service recommendation systems often face two challenges. First, developers are required to input keywords for service search, but due to their lack of knowledge in the relevant field, the keywords entered by the developers are usually freestyle, causing an inability to accurately locate services. Second, it is exceedingly difficult to extract services that meet the requirements due to the 99.8% sparseness of the application service interaction records. To address the above challenges, a framework for service recommendation through multi-model fusion (SRM) is proposed in this paper. Firstly, we employ graph neural network algorithms to deeply mine historical records, extract the features of applications and services, and calculate their preferences. Secondly, we use the BERT model to analyze text information and use the attention mechanism and fully connected neural networks to deeply mine the matching degree between candidate services and development requirements. The two models mentioned above are further merged to obtain the final service recommendation list. Extensive experiments on datasets demonstrate that SRM can significantly enhance the effectiveness of recommendations in service recommendation scenarios.","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"40 12","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139191587","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-01DOI: 10.53106/199115992023123406015
Yan-Ni Zhao Yan-Ni Zhao, Le Xu Yan-Ni Zhao
In the paper, a slicing-guided method is introduced to extract the curve skeleton from the point cloud body model. Firstly, the dominant eigenvector of body model as slicing direction is chosen adaptively, and the input body model is sliced accordingly. each slice is projected and classified into different regions, and the centroid of each region can be considered as initial skeleton point. Then, those skeleton points are removed outside models, and initial skeleton lines are generated by connecting points based on different region of body model. Finally, the two-step post-processing approach is proposed to improve the initial skeleton results for accurate topological analysis. With the branch point merging strategy, the initial skeleton of the model is optimized. Furthermore, the skeleton lines by interpolation optimization are refined and smoothed. Compared with similar skeleton extraction algorithms, the method proposed in the paper has relatively strong robustness and effectiveness, and can be applied to human body model in point cloud data.
{"title":"Slicing-guided Skeleton Extraction Method for 3D Point Clouds of Human Body","authors":"Yan-Ni Zhao Yan-Ni Zhao, Le Xu Yan-Ni Zhao","doi":"10.53106/199115992023123406015","DOIUrl":"https://doi.org/10.53106/199115992023123406015","url":null,"abstract":"In the paper, a slicing-guided method is introduced to extract the curve skeleton from the point cloud body model. Firstly, the dominant eigenvector of body model as slicing direction is chosen adaptively, and the input body model is sliced accordingly. each slice is projected and classified into different regions, and the centroid of each region can be considered as initial skeleton point. Then, those skeleton points are removed outside models, and initial skeleton lines are generated by connecting points based on different region of body model. Finally, the two-step post-processing approach is proposed to improve the initial skeleton results for accurate topological analysis. With the branch point merging strategy, the initial skeleton of the model is optimized. Furthermore, the skeleton lines by interpolation optimization are refined and smoothed. Compared with similar skeleton extraction algorithms, the method proposed in the paper has relatively strong robustness and effectiveness, and can be applied to human body model in point cloud data.","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"5 10","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139189208","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-01DOI: 10.53106/199115992023123406004
Chuqiao Lin Chuqiao Lin, Haoran Sun Chuqiao Lin, Shengda Wang Haoran Sun, Chunyan An Shengda Wang, Han Qi Chunyan An, Xin Luo Han Qi
With the rapid development of information and communication technologies, new services and new applications continue to emerge, especially in smart grid network scenarios. Placing services in terms of containers on the network edge is a promising solution for guaranteeing low latency, large connections, and high bandwidth. In this paper, we propose a multi-objective container migration strategy (MOCMS). Firstly, the container needed to be migrated is selected according to the resource utilization and the energy consumption condition at the network edge. Secondly, in order to avoid the problem of resource fragmentation, the node coordination matrix model is established. Thirdly, in order to obtain the optimal container migration results, an improved Binary Grey Wolf Optimizer (BGWO) algorithm is designed. Finally, the simulation results show that the proposed container migration strategy can perform better than other existing schemes.
{"title":"Container Migration Strategy Based on Multi-objective Optimization for Edge-Cloud Coordination enabled Smart Grids","authors":"Chuqiao Lin Chuqiao Lin, Haoran Sun Chuqiao Lin, Shengda Wang Haoran Sun, Chunyan An Shengda Wang, Han Qi Chunyan An, Xin Luo Han Qi","doi":"10.53106/199115992023123406004","DOIUrl":"https://doi.org/10.53106/199115992023123406004","url":null,"abstract":"With the rapid development of information and communication technologies, new services and new applications continue to emerge, especially in smart grid network scenarios. Placing services in terms of containers on the network edge is a promising solution for guaranteeing low latency, large connections, and high bandwidth. In this paper, we propose a multi-objective container migration strategy (MOCMS). Firstly, the container needed to be migrated is selected according to the resource utilization and the energy consumption condition at the network edge. Secondly, in order to avoid the problem of resource fragmentation, the node coordination matrix model is established. Thirdly, in order to obtain the optimal container migration results, an improved Binary Grey Wolf Optimizer (BGWO) algorithm is designed. Finally, the simulation results show that the proposed container migration strategy can perform better than other existing schemes.","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"24 5","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139191140","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-01DOI: 10.53106/199115992023123406011
Qian-Han Zhang Qian-Han Zhang, Bing-Yan Wei Qian-Han Zhang, Dong-Liang Fan Bing-Yan Wei, Xiao-Ying Wu Dong-Liang Fan, Jin-Ping Du Xiao-Ying Wu
This article focuses on the problems of imperfect models and slow convergence speed of optimization algorithms in the use of photovoltaic microgrids. Firstly, accurate mathematical models are established based on the composition of photovoltaic microgrids, namely photovoltaic power generation systems and energy storage systems. Then, an improved cat swarm algorithm is used to solve the model, ultimately achieving an increase in solving speed during the process while avoiding the algorithm from falling into local optima. Finally, an intelligent monitoring system for photovoltaic microgrids was designed based on the algorithm process of the article, visualizing the main parameters.
{"title":"Design of An Intelligent Monitoring and Control System for Photovoltaic Microgrids","authors":"Qian-Han Zhang Qian-Han Zhang, Bing-Yan Wei Qian-Han Zhang, Dong-Liang Fan Bing-Yan Wei, Xiao-Ying Wu Dong-Liang Fan, Jin-Ping Du Xiao-Ying Wu","doi":"10.53106/199115992023123406011","DOIUrl":"https://doi.org/10.53106/199115992023123406011","url":null,"abstract":"This article focuses on the problems of imperfect models and slow convergence speed of optimization algorithms in the use of photovoltaic microgrids. Firstly, accurate mathematical models are established based on the composition of photovoltaic microgrids, namely photovoltaic power generation systems and energy storage systems. Then, an improved cat swarm algorithm is used to solve the model, ultimately achieving an increase in solving speed during the process while avoiding the algorithm from falling into local optima. Finally, an intelligent monitoring system for photovoltaic microgrids was designed based on the algorithm process of the article, visualizing the main parameters.","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"22 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139191303","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-12-01DOI: 10.53106/199115992023123406001
Jin Dai Jin Dai, Zhifang Zheng Jin Dai
Variational autoencoder (VAE) has the problem of uninterpretable data generation process, because the features contained in the VAE latent space are coupled with each other and no mapping from the latent space to the semantic space is established. However, most existing algorithms cannot understand the data distribution features in the latent space semantically. In this paper, we propose a cloud model-based method for disentangling semantic features in VAE latent space by adding support vector machines (SVM) to feature transformations of latent variables, and we propose to use the cloud model to measure the degree of disentangling of semantic features in the latent space. The experimental results on the CelebA dataset show that the method obtains a good disentangling effect of semantic features in the latent space, which proves the effectiveness of the method from both qualitative and quantitative aspects.
{"title":"Disentangling Representation of Variational Autoencoders Based on Cloud Models","authors":"Jin Dai Jin Dai, Zhifang Zheng Jin Dai","doi":"10.53106/199115992023123406001","DOIUrl":"https://doi.org/10.53106/199115992023123406001","url":null,"abstract":"Variational autoencoder (VAE) has the problem of uninterpretable data generation process, because the features contained in the VAE latent space are coupled with each other and no mapping from the latent space to the semantic space is established. However, most existing algorithms cannot understand the data distribution features in the latent space semantically. In this paper, we propose a cloud model-based method for disentangling semantic features in VAE latent space by adding support vector machines (SVM) to feature transformations of latent variables, and we propose to use the cloud model to measure the degree of disentangling of semantic features in the latent space. The experimental results on the CelebA dataset show that the method obtains a good disentangling effect of semantic features in the latent space, which proves the effectiveness of the method from both qualitative and quantitative aspects.","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"85 5","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139192714","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The paper develops a set of mobile laboratories based on the grid data cloud platform. The laboratory proposed a multi-data flow cooperative algorithm based on a cross-bus four-layer temporal space model and a cross-directional multiplier. This algorithm achieves the purpose of updating most data streams as a whole. The system establishes a statistical analysis system of data flow from multiple perspectives and mines and monitors multiple data flows. The paper divides most data streams into several linear modules, and corresponding matrices are formed. Finally, the simulated test results of the paper show that the CPU usage of the mobile laboratory computer-aided system based on the power network is very small. The system processing efficiency is high.
{"title":"Research on Computer Aided Laboratory Based on Cross-Multiplier Multi-Data Flow Collaborative Algorithm","authors":"Hao Wu Hao Wu, Xiao Xu Hao Wu, Ninghui Guo Xiao Xu, Zinan Peng Ninghui Guo, Yujia Zhai Zinan Peng, Sijia Wu Yujia Zhai","doi":"10.53106/199115992023083404007","DOIUrl":"https://doi.org/10.53106/199115992023083404007","url":null,"abstract":"\u0000 The paper develops a set of mobile laboratories based on the grid data cloud platform. The laboratory proposed a multi-data flow cooperative algorithm based on a cross-bus four-layer temporal space model and a cross-directional multiplier. This algorithm achieves the purpose of updating most data streams as a whole. The system establishes a statistical analysis system of data flow from multiple perspectives and mines and monitors multiple data flows. The paper divides most data streams into several linear modules, and corresponding matrices are formed. Finally, the simulated test results of the paper show that the CPU usage of the mobile laboratory computer-aided system based on the power network is very small. The system processing efficiency is high.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124971503","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Crack detection is an important aspect to measure the structural stability of buildings. At present, the detection of building cracks still mainly adopts manual detection methods, which rely too much on personal experience, low detection accuracy, and consume a lot of manpower and material resources. In response to this issue, we use an end-to-end method to predict the pixel by pixel crack segmentation DeepCrack network model, and use CRF and GF methods to fuse the final prediction results. Firstly, the ResNet34 model was pre trained on the PASCAL VOC2007 dataset. The DeepCrack + CRF + GF model was used for training, and the Adaptive Threshold method was used to partition and binarize the training results. Finally, the constructed wall crack detection model achieved an AP value of 89.12%, accuracy and recall rates of 83.96%, 88.47%, and IoU value of 85.80%. On the premise of ensuring detection accuracy, the model is only 47 MB, making it possible to deploy it on embedded devices. It can be used in practical engineering applications to build an intelligent building crack detection system, saving a lot of manpower and resources.
{"title":"Intelligent Crack Detection and Analysis of Building Walls Based on DeepCrack Network","authors":"Yinggang Xie Yinggang Xie, XueWei Peng YingGang Xie, YangPeng Xiao XueWei Peng, YaRu Zhang YangPeng Xiao","doi":"10.53106/199115992023083404018","DOIUrl":"https://doi.org/10.53106/199115992023083404018","url":null,"abstract":"\u0000 Crack detection is an important aspect to measure the structural stability of buildings. At present, the detection of building cracks still mainly adopts manual detection methods, which rely too much on personal experience, low detection accuracy, and consume a lot of manpower and material resources. In response to this issue, we use an end-to-end method to predict the pixel by pixel crack segmentation DeepCrack network model, and use CRF and GF methods to fuse the final prediction results. Firstly, the ResNet34 model was pre trained on the PASCAL VOC2007 dataset. The DeepCrack + CRF + GF model was used for training, and the Adaptive Threshold method was used to partition and binarize the training results. Finally, the constructed wall crack detection model achieved an AP value of 89.12%, accuracy and recall rates of 83.96%, 88.47%, and IoU value of 85.80%. On the premise of ensuring detection accuracy, the model is only 47 MB, making it possible to deploy it on embedded devices. It can be used in practical engineering applications to build an intelligent building crack detection system, saving a lot of manpower and resources.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126206277","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.53106/199115992023083404021
Xiaoxin Guo Xiaoxin Guo, Xintai Liu Xiaoxin Guo, Haixia Liu Xintai Liu
In order to make the railroad inspection robot better adapt to its complex working environment, it is especially important to study the robot object avoidance algorithm. The WOA algorithm has simple and understandable structure and strong optimization ability but is prone to local convergence. IWOA-PSO is used for railway inspection robots. The performance of IWOA-PSO in the experimental results is better than that of WOA and PSO, and the average accuracy and standard deviation of the IWOA-PSO can better reach the theoretical optimal value in the function tests, and it has performance close to the theoretical value. In the simple environment object avoidance route planning, the minimum path length of IWOA-PSO is 850 mm, which is 53.6% less than that of the PSO algorithm, and the search time is 13.12 seconds, which is 5.11 seconds less than that of PSO algorithm; in the ordinary environment object avoidance route planning, the minimum path length of IWOA-PSO is 830 mm, while the path length of PSO algorithm is 1339 mm, the former is 38% less than the latter, and the search time of IWOA-PSO is 14.05 seconds less than PSO algorithm, so the method has better effect on object avoidance.
{"title":"Intelligent Object Avoidance Method Design of Railroad Inspection Robot Based on Particle Swarm Algorithm","authors":"Xiaoxin Guo Xiaoxin Guo, Xintai Liu Xiaoxin Guo, Haixia Liu Xintai Liu","doi":"10.53106/199115992023083404021","DOIUrl":"https://doi.org/10.53106/199115992023083404021","url":null,"abstract":"\u0000 In order to make the railroad inspection robot better adapt to its complex working environment, it is especially important to study the robot object avoidance algorithm. The WOA algorithm has simple and understandable structure and strong optimization ability but is prone to local convergence. IWOA-PSO is used for railway inspection robots. The performance of IWOA-PSO in the experimental results is better than that of WOA and PSO, and the average accuracy and standard deviation of the IWOA-PSO can better reach the theoretical optimal value in the function tests, and it has performance close to the theoretical value. In the simple environment object avoidance route planning, the minimum path length of IWOA-PSO is 850 mm, which is 53.6% less than that of the PSO algorithm, and the search time is 13.12 seconds, which is 5.11 seconds less than that of PSO algorithm; in the ordinary environment object avoidance route planning, the minimum path length of IWOA-PSO is 830 mm, while the path length of PSO algorithm is 1339 mm, the former is 38% less than the latter, and the search time of IWOA-PSO is 14.05 seconds less than PSO algorithm, so the method has better effect on object avoidance.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115932759","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.53106/199115992023083404022
Xin-Yi Wang Xin-Yi Wang, Hao-Ran Sun Xin-Yi Wang, Xu-Yang Yin Hao-Ran Sun, Chun-Zi Li Xu-Yang Yin, Sheng-Yu Liu Chun-Zi Li
Collaborative filtering-based models can use the interaction between users and products or the correlation between users and users, and between products and products. However, methods based on collaborative filtering can only grasp one type of relationship and still cannot fully fit. Various factors influencing user preferences make a lot of redundant information still not filtered out. We proposals a collaborative filtering model based on deep learning, which combines the item-item relationship learning in advance with a neural collaborative filtering network to effectually make recommendations. In the initial stage, learn low-dimensional vectors of compartments, and embed information that reflections the co-occurrence relationship between compartments. The prediction stage combines the trained embedding vector with the embedding vector of the module as a correction to the output result of the neural network. The benchmark data set MovieLens 1M is the experienced data set of this article, and the effectiveness of this method is verified on the data set. The experienced results are compared with some advanced methods on the data set. The results show that the model proposed in this paper is better than some methods based on collaborative filtering.
{"title":"Deep Collaborative Filtering System","authors":"Xin-Yi Wang Xin-Yi Wang, Hao-Ran Sun Xin-Yi Wang, Xu-Yang Yin Hao-Ran Sun, Chun-Zi Li Xu-Yang Yin, Sheng-Yu Liu Chun-Zi Li","doi":"10.53106/199115992023083404022","DOIUrl":"https://doi.org/10.53106/199115992023083404022","url":null,"abstract":"\u0000 Collaborative filtering-based models can use the interaction between users and products or the correlation between users and users, and between products and products. However, methods based on collaborative filtering can only grasp one type of relationship and still cannot fully fit. Various factors influencing user preferences make a lot of redundant information still not filtered out. We proposals a collaborative filtering model based on deep learning, which combines the item-item relationship learning in advance with a neural collaborative filtering network to effectually make recommendations. In the initial stage, learn low-dimensional vectors of compartments, and embed information that reflections the co-occurrence relationship between compartments. The prediction stage combines the trained embedding vector with the embedding vector of the module as a correction to the output result of the neural network. The benchmark data set MovieLens 1M is the experienced data set of this article, and the effectiveness of this method is verified on the data set. The experienced results are compared with some advanced methods on the data set. The results show that the model proposed in this paper is better than some methods based on collaborative filtering.\u0000 \u0000","PeriodicalId":345067,"journal":{"name":"電腦學刊","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130186142","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：