A Recommendation Algorithm Based on Efficient Dense Subsequence and Data Generation

As an efficient information filtering tool, the recommendation system can help users obtain valuable information quickly and efficiently while effectively alleviating the information load problem. However, the classic collaborative filtering algorithm has data sparsity problems in practical applications, which affects the accuracy of recommendation. To solve this problem, this paper took movie recommendation as the research object and proposed a hybrid recommendation algorithm named K-GBDT which combined KNN and gradient boosting decision tree. Firstly, it used KNN to obtain the similar information of the target user and item for preliminary prediction. At the same time, it mined users' basic information features and potential features of users and items from the original data. Then, it tried to adopt XGBoost, LightGBM and CatBoost algorithm which were proposed base the idea of gradient boosting decision tree to build regression model. Finally, the target user's rating for the movie was predicted with the trained model, and the recommendation list was generated based on the predicted results. Compared with some classic recommendation algorithms such as NMF, Slope One, Co-Clustering and etc. Based on MovieLens datasets, results show that the mean absolute error and root mean square error of this algorithm are both low, and its recommendation accuracy is higher.

Introduction: Recommendation systems (RS) are pivotal in enhancing personalized user experiences, particularly in the educational domain, where tailored book recommendation systems (BRS) support individualized learning. Despite their importance, selecting an optimal recommendation algorithm remains challenging, especially in scenarios involving large and sparse datasets. Objectives: This study aims to evaluate the performance of four recommendation algorithms—Random Baseline, Popular, User-Based Collaborative Filtering (UBCF), and Singular Value Decomposition (SVD)—to identify their strengths and tradeoffs regarding accuracy and computational efficiency. The goal is to provide actionable insights for selecting appropriate algorithms for educational platforms and similar applications. Methods: The evaluation employed a dataset containing over 10,000 books, 53,424 users, and 981,756 ratings. The algorithms were assessed using Root Mean Square Error (RMSE) to measure predictive accuracy and training time to indicate computational efficiency. The analysis addressed challenges such as data sparsity and rating biases. Results: The results demonstrated that SVD achieved the highest accuracy (RMSE: 0.950), effectively uncovering latent relationships in sparse datasets. UBCF, with a slightly lower accuracy (RMSE: 1.020), offered a balance between accuracy and computational efficiency, making it suitable for real-time applications. Conversely, simpler algorithms like Random Baseline and Popular exhibited faster training times but significantly lower predictive accuracy, highlighting the limitations of non-personalized methods. Conclusions: This study underscores the tradeoffs between accuracy and efficiency among recommendation algorithms. While SVD is ideal for accuracy-driven applications, UBCF provides a practical alternative for scenarios requiring computational efficiency. The findings have substantial implications for educational platforms and e-commerce, where personalized recommendations enhance user satisfaction and engagement. Future research should focus on integrating deep learning models and expanding evaluation criteria, including user satisfaction and diversity, to further improve the performance of recommendation systems.

The increasing demand for personalized information has resulted in the development of the Recommender System (RS). RS has been widely utilized and broadly studied to suggest the interests of users and make an appropriate recommendation. This paper gives an overview of several types of recommendation approaches based on user preferences, ratings, domain knowledge, users demographic data, users context and also lists the advantages and disadvantages of each RS approach. In this paper, we also proposed the movie recommendation based on collaborative filtering and singular value decomposition plus-plus (SVD++). The proposed approach is compared with well-known machine learning approaches namely k nearest neighbor (K-NN), singular value decomposition (SVD) and Co-clustering. The proposed approach is experimentally verified using MovieLens 100 K datasets and error of the RS is measured using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). The result shows that the proposed approach gives a lesser error rate with RMSE (0.9201) and MAE (0.7219). This approach also overcomes cold-start, data sparsity problems and provides them relevant items and services.

The successful design of collaborative filtering system for recommending depends hardly on finding the nearest neighbors. In this paper, we provide a new collaborative filtering method based on concept lattices to generate more precise recommendation. Firstly, we analyze the log files and construct a formal context which is then used to build a concept lattice. Based on the concept lattice, we propose a new way of dynamically recognizing the user's independent interests and making a Top-N recommendation to suggest what the user may want to visit next. One advantage of our method is that it can quickly group user's interest based on their navigation history. Besides, a new algorithm based on concept lattice is proposed ground on multi-attribute decision making theory. Further more, it smoothed the cold start problem in traditional collaborative filtering methods. The developed method is validated by the practical web access logs. The experiment shows our approach is practical, feasible and efficient.

Recommendation systems have been widely used in many internet communities, as they could provide users with personalized items based on the big data analysis technologies. Collaborative Filtering (CF) recommendation algorithm is a well-known and widely used technique in the recommendation systems. However, traditional collaborative filtering methods usually make predictions only based on the rating-based features, which could not make a better understanding of user preferences. Moreover, with the growing disclosure of online personal information, privacy protection is another important issue to be considered in dynamic cyber networks. Thus, a security-aware prediction mechanism is necessary to deal with the data privacy in recommendation systems. In view of these challenges, this paper proposes a privacy-aware multi-preference-based collaborative filtering recommendation method with Locality-Sensitive Hashing (LSH). Firstly, two types of user representations are obtained based on both rating features and latent factors. To protect the privacy data of users, both the LSH technique and differential privacy data perturbation mechanism are integrated into our recommendation model. Besides, the employment of LSH in establishing users' nearest neighbor matrix also improve the similarity searching efficiency of the CF recommendation algorithm. Final recommendation is based on a SGD-based hybrid collaborative prediction model. Experimental results show that our proposed algorithm can improve the accuracy of the recommendation system and reduce the time cost on the basis of ensuring user privacy.

INTRODUCTION: Collaborative filtering (CF) algorithms based on deep learning and matrix factorization aim to learn deeper latent features of users and items from user ratings. However, such methods often face challenges due to the sparsity of rating data, which limits their recommendation performance. OBJECTIVES: This study aims to design a hybrid recommendation algorithm named auto-encoder deep learning and matrix factorization (AED-MF), which integrates deep learning and matrix factorization to address the sparsity issue in rating data and improve the extraction of deeper latent features of users and items. METHODS: The AED-MF algorithm combines an auto-encoder-based deep learning approach with matrix factorization to model both deeper hidden representations and nonlinear relationships between users and items. The methodology includes data downloading, mounting, cleaning, model training, and experimental evaluation, with matrix factorization applied to predict missing ratings in the rating matrix. RESULTS: The proposed AED-MF algorithm demonstrated strong performance in key recommendation metrics, effectively learning deeper user and item representations and applying them to capture complex nonlinear relationships. It also reduced the impact of rating data sparsity while maintaining high recommendation accuracy. CONCLUSION: The AED-MF recommendation algorithm successfully alleviates the problem of sparse rating data and enhances recommendation accuracy by leveraging the strengths of both deep learning and matrix factorization, offering an effective solution for improving collaborative filtering-based recommender systems.

Traditional collaborative filtering algorithm does not take into account the user’s interest factors, at the same time, it has the problem of sparse data, poor scalability and so on, which directly affects the quality of the recommendation. This paper proposes a collaborative filtering recommendation algorithm based on user interest clustering. Firstly, according to the user’s existing score, a method is used to compute the interest degree of the user. Secondly, a clustering algorithm is used to divide the users into clusters. Finally, a collaborative filtering algorithm is used to recommend movie, it effectively improves data sparsity and real-time of problem. The experimental results exhibit that the proposed algorithm shows a great improvement in the recommendation accuracy.

Social recommendation leverages social information to solve data sparsity and cold-start problems in traditional collaborative filtering methods. However, most existing models assume that social effects from friend users are static and under the forms of constant weights or fixed constraints. To relax this strong assumption, in this paper, we propose dual graph attention networks to collaboratively learn representations for two-fold social effects, where one is modeled by a user-specific attention weight and the other is modeled by a dynamic and context-aware attention weight. We also extend the social effects in user domain to item domain, so that information from related items can be leveraged to further alleviate the data sparsity problem. Furthermore, considering that different social effects in two domains could interact with each other and jointly influence user preferences for items, we propose a new policy-based fusion strategy based on contextual multi-armed bandit to weigh interactions of various social effects. Experiments on one benchmark dataset and a commercial dataset verify the efficacy of the key components in our model. The results show that our model achieves great improvement for recommendation accuracy compared with other state-of-the-art social recommendation methods.

The most classic algorithm in the recommender system domain is the collaborative filtering algorithm, but it often suffers from data sparsity and item cold-start problems. Moreover, most traditional collaborative filtering algorithms only consider the user's rating of the item, which greatly limits the accuracy of the recommendation. This paper takes movie recommendation as the research object, and proposes a collaborative filtering recommendation model based on LDA topic model and network embedding, which improves the dependability of the model. The model first uses the LDA topic model to extract the text features of users and movies from texts, then builds the association network through the association features between users, and learns the network features using node2vec and GraphGAN. Finally, the two parts are combined to compute the similarity between users to recommend movies using the collaborative filtering algorithm. Results of the experiment on the real dataset of the Douban website demonstrate that movie recommendation with the user characteristics mined by the LDA model and GraphGAN combined can achieve the best effect, which is much higher than the collaborative filtering based on user ratings or using only LDA theme. This proves that network embedding can promote the accuracy of recommendationation, while ensuring the dependability of the model. The method proposed in this paper is effective.

The research aims to solve the challenge of filtering out the content that users are really interested in from massive data. In this paper, an intelligent recommendation system named multi-source fusion deep collaborative recommendation algorithm (MSDC-Rec) is proposed, which combines collaborative filtering, content filtering and deep learning (DL) technology to improve the accuracy and diversity of the recommendation system. MSDC-Rec algorithm integrates user behavior data, item attribute data and social network information, and uses deep neural network for advanced feature extraction and pattern learning, thus accurately portraying user portraits and realizing personalized and accurate recommendation services. The experimental results show that MSDC-Rec is superior to the traditional algorithm in accuracy, recall, F1 score and personalized index, and has low root mean square error (RMSE) and acceptable real-time. This study not only solves the problems of data sparsity and cold start, but also significantly improves the accuracy and personalization of recommendation, which provides a new idea for the development of intelligent recommendation system.

Background and purposeThe study evaluated the differences in leaf positioning deviations by the log files of three advanced accelerators with two delivery techniques, and established specific assessment parameters of leaf positioning deviations for different types of accelerators.MethodsA total of 420 treatment plans with 5 consecutive treatment log files were collected from the Trilogy, TrueBeam and Halcyon accelerators. Millennium MLC was equipped on the Trilogy and TrueBeam accelerators. A jawless design and dual-layer MLC were adopted on the Halcyon accelerator. 70 IMRT and 70 VMAT plans were selected randomly on each accelerator. The treatment sites of all plans included head and neck, chest, breast, pelvis and other sites. The parsing tasks for 2100 log files were proceeded by SunCheck software from Sun Nuclear Corporation. The maximum leaf root mean square (RMS) errors, 95th percentile errors and percentages of different leaf positioning errors were statistically analyzed. The correlations between these evaluation parameters and accelerator performance parameters (maximum leaf speed, mean leaf speed, gantry and arc angle) were analyzed.ResultsThe average maximum leaf RMS errors of the Trilogy in the IMRT and VMAT plans were 0.44 ± 0.09 mm and 0.79 ± 0.07 mm, respectively, which were higher than the TrueBeam's 0.03 ± 0.01 mm, 0.03 ± 0.01 mm and the Halcyon's 0.05 ± 0.01 mm, 0.07 ± 0.01 mm. Similar data results were shown in the 95th percentile error. The maximum leaf RMS errors were strongly correlated with the 95th percentile errors (Pearson index > 0.5). The leaf positioning deviations in VMAT were higher than those in IMRT for all accelerators. In TrueBeam and Halcyon, leaf position errors above 1 mm were not found in IMRT and VMAT plans. The main influencing factor of leaf positioning deviation was the leaf speed, which has no strong correlation with gantry and arc angles.ConclusionsCompared with the quality assurance guidelines, the MLC positioning deviations tolerances of the three accelerators should be tightened. For both IMRT and VMAT techniques, the 95th percentile error and the maximum RMS error are suggested to be tightened to 1.5 and 1 mm respectively for the Trilogy accelerator. In TrueBeam and Halcyon accelerators, the 95th percentile error and maximum RMS error of 1 and 0.5 mm, respectively, are considered appropriate.

BackgroundThe multileaf collimator (MLC) is a critical component to accurate intensity-modulated radiotherapy (IMRT) delivery. This study examined MLC positional accuracy via MLC logs from Varian machines from six institutions and three delivery techniques to evaluate typical positional accuracy and treatment and mechanical parameters that affect accuracy. Typical accuracy achieved was compared against TG-142 recommendations for MLC performance; more appropriate recommendations are suggested.MethodsOver 85,000 Varian MLC treatment logs were collected from six institutions and analyzed with FractionCHECK. Data were binned according to institution and treatment type to determine overall root mean square (RMS) and 95th percentile error values, and then to look for correlations between those errors and with mechanical and treatment parameters including mean and maximum leaf speed, gantry angle, beam-on time, mean leaf error, and number of segments.ResultsResults of treatment logs found that leaf RMS error and 95th percentile leaf error were consistent between institutions, but varied by treatment type. The step and shoot technique had very small errors: the mean RMS leaf error was 0.008 mm. For dynamic treatments the mean RMS leaf error was 0.32 mm, while volumetric-modulated arc treatment (VMAT) showed an RMS leaf error of 0.46 mm. Most MLC leaf errors were found to be well below TG-142 recommended tolerances. For the dynamic and VMAT techniques, the mean and maximum leaf speeds were significantly linked to the leaf RMS error. Additionally, for dynamic delivery, the mean leaf error was correlated with RMS error, whereas for VMAT the average gantry speed was correlated. For all treatments, the RMS error and the 95th percentile leaf error were correlated.ConclusionsRestricting the maximum leaf speed can help improve MLC performance for dynamic and VMAT deliveries. Furthermore, the tolerances of leaf RMS and error counts for all treatment types should be tightened from the TG-142 values to make them more appropriate for clinical performance. Values of 1 mm for the 95th percentile of leaf RMS error and 1.5 mm for the 95th percentile leaf error are suggested as action levels for all treatment types.

SummaryWith the advent of Internet of Things (IoT) age, the variety and volume of web services have been increasing at a fast speed. This often leads to users' selections for web services more complicated. Under the circumstance, a variety of methods such as collaborative filtering are adopted to deal with this challenging situation. While traditional collaborative filtering method has some shortcomings, one of which is that only centralized user‐service data are considered while distributed quality data from multiple platform are ignored. Generally, service recommendation across different platforms often involves data communication among multiple platforms, during which user privacy may be disclosed and much computational time is required. Considering these challenges, a unique amplified locality‐sensitive hashing (LSH)‐based service recommendation method, that is, SRAmplified‐LSH, is proposed in the article. SRAmplified‐LSH can guarantee a good balance between accuracy and efficiency of recommendation and user privacy information. Finally, extensive experiments deployed on WS‐DREAM dataset validate the feasibility of our proposed method.

This study proposes a hybrid recommendation framework for online shopping platforms that integrates Convolutional Neural Networks (CNNs) with collaborative filtering (CF) to improve personalization accuracy. Traditional recommendation approaches, such as content-based and collaborative filtering methods, often suffer from data sparsity, cold-start problems, and limited capability to capture rich item features, which reduces recommendation effectiveness in large-scale e-commerce systems. To overcome these challenges, the proposed model leverages CNNs to extract visual and textual features of products while CF learns latent user–item interaction patterns. By combining content-level representations with hidden user preferences and engagement behavior, the hybrid approach enhances recommendation relevance and accuracy. The model was trained and evaluated using the Kaggle E-Commerce Product Recommendation Collaborative dataset. Experimental results demonstrate superior prediction performance compared to conventional methods, highlighting improved scalability, accuracy, and user engagement. The findings confirm the effectiveness of the CNN-based collaborative filtering framework as a robust and data-driven solution for modern e-commerce recommendation systems, enabling more personalized and relevant shopping experiences.

This significant article focuses on the product recommendation related issues involved in the entire process of power marketing. It takes the user collaborative filtering recommendation algorithm and project collaborative filtering algorithm as a solid foundation, and creatively proposes a collaborative filtering hybrid recommendation optimization model. By fully utilizing various means to deeply mine the historical data of users, we have successfully obtained their interests and preferences, and meticulously classified and divided different user interests and preferences, ultimately being able to accurately recommend various types of items that meet their preferences to users. A large number of experimental results fully demonstrate that the carefully proposed recommendation algorithm in this article has a significantly higher level of recommendation accuracy, and its algorithm performance also shows better performance.