top-N Recommendation Research Articles

Product Personalized Recommendation Engine Based on Collaborative Filtering Algorithm

Nowadays, information plus Internet technology is continuously developing, and people have entered the era of information overload from the era of lack of information access. In this context, it is difficult to find one's own "tailor-made" information from the huge amount of information. In this context, the recommendation system came into being. This article uses the order data under an e-commerce platform, uses user-based and item-based collaborative filtering algorithms to personalize products, and builds a product recommendation system. Through the feasibility of actual application scenarios and the prediction score of the model, a series of evaluations were made on the algorithm, and reliable and effective prediction results were obtained from the three indicators of accuracy, recall and F1 score. The final recommendation effect of the TopN recommendation list obtained was significant, and the recommendation efficiency could reach 85%.Provide some ideas for the research and development direction of modern recommendation algorithms.

A Revisiting Study of Appropriate Offline Evaluation for Top- N Recommendation Algorithms

In recommender systems, top- N recommendation is an important task with implicit feedback data. Although the recent success of deep learning largely pushes forward the research on top- N recommendation, there are increasing concerns on appropriate evaluation of recommendation algorithms. It therefore is important to study how recommendation algorithms can be reliably evaluated and thoroughly verified. This work presents a large-scale, systematic study on six important factors from three aspects for evaluating recommender systems. We carefully select 12 top- N recommendation algorithms and eight recommendation datasets. Our experiments are carefully designed and extensively conducted with these algorithms and datasets. In particular, all the experiments in our work are implemented based on an open sourced recommendation library, Recbole [ 139 ], which ensures the reproducibility and reliability of our results. Based on the large-scale experiments and detailed analysis, we derive several key findings on the experimental settings for evaluating recommender systems. Our findings show that some settings can lead to substantial or significant differences in performance ranking of the compared algorithms. In response to recent evaluation concerns, we also provide several suggested settings that are specially important for performance comparison.

Methods for explaining Top-N recommendations through subgroup discovery

Methods for explaining Top-N recommendations through subgroup discovery

Application of Internet of Things technology in Human resource recommendation engine

In order to avoid the phenomenon of information overload in the field of human resources, the author proposes a human resource recommendation engine based on Internet of Things technology. Firstly, the application fields of the Internet of Things are introduced, and the human resource recommendation engine is designed, use a mixed referral approach in the HR field. We chose a hybrid recommendation strategy of PLSA and content based on the Internet of Things. The author introduces the human resource recommendation and content-based human resource recommendation of PLSA based on the Internet of Things respectively, on the basis of the above introduction, through the analysis of the existing hybrid recommendation strategy, a weighted hybrid method of PLSA based on the Internet of Things and content-based recommendation algorithm is proposed. In order to verify the feasibility of the proposed IoT-based PLSA and content-based HR recommendation method, we conducted a multi-step experiment. We selected 50 graduating students from A University to participate in this experiment, and the results show that in order to verify the overall effect of the algorithm, with accuracy and recall as evaluation criteria, we conducted experiments on recommendation algorithms of PLSA based on Internet of Things, content-based recommendation algorithms, and recommendation algorithms based on PLSA and content weighting mixture, where, the number of z in PLSA algorithm is 10, the value of β in mixed recommendation is 0.6, the size of N in TopN recommendation is N=5,10,20, and 30. Experiments were carried out on the accuracy and recall of the three recommendation algorithms, the number of z in PLSA algorithm is 10, the value of β is 0.6, and the size of N in TopN recommendation is N=3,5,7,9. The recommendation effect of PLSA based on the Internet of Things is better than that of content-based recommendation algorithm, and the recommendation effect of the combination of the two algorithms is better than either of them.

Fully adaptive recommendation paradigm: top-enhanced recommender distillation for intelligent education systems

Top-N recommendation has received great attention in assisting students in providing personalized learning guidance on the required subject/domain. Generally, existing approaches mainly aim to maximize the overall accuracy of the recommendation list while ignoring the accuracy of highly ranked recommended exercises, which seriously affects the students’ learning enthusiasm. Motivated by the Knowledge Distillation (KD) technique, we skillfully design a fully adaptive recommendation paradigm named Top-enhanced Recommender Distillation framework (TERD) to improve the recommendation effect of the top positions. Specifically, the proposed TERD transfers the knowledge of an arbitrary recommender (teacher network), and injects it into a well-designed student network. The prior knowledge provided by the teacher network, including student-exercise embeddings, and candidate exercise subsets, are further utilized to define the state and action space of the student network (i.e., DDQN). In addition, the student network introduces a well-designed state representation scheme and an effective individual ability tracing model to enhance the recommendation accuracy of top positions. The developed TERD follows a flexible model-agnostic paradigm that not only simplifies the action space of the student network, but also promotes the recommendation accuracy of the top position, thus enhancing the students’ motivation and engagement in e-learning environment. We implement our proposed approach on three well-established datasets and evaluate its Top-enhanced performance. The experimental evaluation on three publicly available datasets shows that our proposed TERD scheme effectively resolves the Top-enhanced recommendation issue.

Correction to: Top-N Recommendation System Using Explicit Feedback and Outer Product Based Residual CNN

Correction to: Top-N Recommendation System Using Explicit Feedback and Outer Product Based Residual CNN

Selection bias mitigation in recommender system using uninteresting items based on temporal visibility

Most collaborative filtering recommendation algorithms rely too much on the user's historical rating data. However, selection bias is common in explicit feedback data, which makes the learning of user preferences face more challenges. We verify the influence of selection bias on topN recommendation, and propose a data filling strategy using uninteresting items based on temporal visibility to alleviate the selection bias in the data. Specifically, our method includes a weighted matrix factorization model to learn users' pre-use preferences for unrated items. According to the experience of items that users have seen but not interacted show negative preferences, we combine user activity, item popularity and temporal rating information to carry out non-uniform weighting to evaluate the confidence of unrated items as a negative example. Then the items with low pre-use preferences are taken as uninteresting items and filled in a low value to restore the user's real rating distribution. Experiments on two real world datasets show that our algorithm can effectively alleviate the selection bias and improve the recommendation accuracy.

Trust-aware denoising autoencoder with spatial-temporal activity for cross-domain personalized recommendations

Recently, cross-domain recommendation systems have been very helpful in improving the quality of recommendation and solving the problem of cold start and data sparsity. Cross-domain recommender systems allow the transfer of knowledge from one domain with dense ratings to other domain with sparse ratings. Such transfer of knowledge helps in addressing the data sparseness and cold start issues in traditional recommender systems. Although cross-domain recommendations have evolved significantly, yet employment of various factors, such as time, trust, and location remains a challenge. Most of the existing approaches ignore the important fact that at what specific time the user may be interested in the recommended item. Moreover, a person’s trust level and sentiments may be influenced by the variation in the location and time, thereby affecting the decision making. In this paper, we propose a cross-domain recommender system that not only takes into account the time at finer granularity levels (e.g., hours, days, weeks, etc.), but also considers a persons location, trust level, and sentiment analysis while computing recommendations. Our proposed model, named as, Trust-Aware Spatial-Temporal Activity based Denoising Autoencoder (TSTDAE), employs autoencoder-based deep-learning models to generate top-N recommendations for a given user and addresses the cold-start problem in the cross-domain scenario of ‘User Overlap’. The proposed work is fivefold: i) Filter out the users’ biased profiles based on sentiment analysis. ii) Learn the features using autoencoder and then perform clustering among the users of source and target domains to discover the best neighbors. iii) Compute the trust and ratio of preference bias between active user (the user to whom top-N items are recommended) and their neighbors and grade the neighbors based on unbiased preferences iv) Project the best time for recommending the items to an active user and generate the top-N recommendations. We have evaluated the proposed model on a public dataset of e-commerce retail service ‘AliExpress’ for the evaluation metrics: Precision, Mean Absolute Precision (MAP), Normalized Discounted Cumulative Gain (NDCG), and Hit Ratio (HR). The experimental results showed improved performance of the proposed system over the existing models.

Enhancing Context-Aware Recommendation Using Hesitant Fuzzy Item Clustering by Stacked Autoencoder Based Smoothing Technique

Context-aware recommender systems (CARS) are a key component in businesses, notably in the e-transactions domain, since they assume that reviews, ratings, demographics, and other factors may determine customer preferences. On contrary, while evaluating the sentiment underlying the reviews and the rating score, consumers’ opinion is typically conflicting. As a result, a framework that employs either a review or a rating for top-N recommendation directs to produce unsatisfied recommendations in addition to a meager rating problem and high computation time. To overcome all the problems, a novel sentiment enhanced stacked autoencoder (SSAE) with context-specific hesitant fuzzy item hierarchical clustering (CHFHC) approach is proposed which employs online and offline phases. In the offline-phase, the meager user-item rating matrix is smoothed by learning the users’ concrete preference to a complete matrix by the SSAE approach. They are clustered offline using the CHFHC approach into context-based similar item clusters. In the online-phase, the active user gets context-based recommendations from the most similar cluster that matches the active users’ current context situation. Hence the SSAE_CHFHC approach improves the quality of Top-N recommendation corresponding to the exact contextual situation of the active user with a minimal recommendation computation time. Experiments on the (5-core) Amazon and yelp datasets proved that the intended SSAE_CHFHC framework consistently outperforms state-of-the-art recommendation algorithms on a variety of evaluation measures.

Unifying attentive sparse autoencoder with neural collaborative filtering for recommendation

The autoencoder network has been proven to be one of the powerful techniques for recommender systems. Currently, the ways of utilizing autoencoder in recommender systems can be divided into two categories: modeling user-item interaction rely solely on autoencoder and integrating autoencoder with other models. Most existing methods based on autoencoder assume that all features of model’s input are equally the same contributing to the final prediction, which can be regarded as attention weight vectors; however, this hypothesis is not reliable, especially when exploring users’ interaction frequency with different items. Moreover, combining autoencoder with traditional methods, the usual strategy is to leverage a linear kernel of the inner product of user and item vectors to predict user preferences, which will lead to insufficient expression power and hurt the performance of recommendation when facing data sparsity and cold start problems. To tackle the above two problems, we propose a novel hybrid deep learning model for top-n recommendation, called attentive stacked sparse autoencoder (A-SAERec), which can capture attention weights vector of a user for items, and then combined with the neural matrix factorization to improve the performance of recommender model. Extensive experiments on four real-world datasets show that our A-SAERec algorithm has significant improvements over state-of-the-art algorithms.

Multi-view hybrid recommendation model based on deep learning

With the rapid development of technologies such as cloud computing, big data, and the Internet of Things, the scale of data continues to grow. The recommendation system has become one of the important intelligent software to help users make decisions. The recommendation model based on user rating data is widely studied and applied, but the data sparsity problem and the cold start problem seriously affect the recommendation quality. In this paper, Multi-view Hybrid Recommendation Model (MHRM) based on deep learning is proposed. First, we use WLDA (an improved Latent Dirichlet Allocation method) to extract the vector representation of user review text, and then apply LSTM to contextual semantic level user review sentiment analysis. At the same time, the emotion fusion method based on user score embedding is proposed. The problems such as deviations between the user score and actual interest preference, and unbalanced distribution of the score level are solved. This paper has been tested on Amazon product data and compared with various classic recommendation algorithms, using Mean Absolute Error (MAE), hit rate and standardized discount cumulative return for performance evaluation. The experimental results show that the prediction of the MHRM proposed in this paper on the 7 recommendation data and the TopN recommendation index have been significantly improved.

An optimized recommendation framework exploiting textual review based opinion mining for generating pleasantly surprising, novel yet relevant recommendations

Serendipity is a critical factor in the Recommender Systems (RS) in delivering pleasantly surprising, novel, yet contextually relevant recommendations. Most existing methods improve serendipity in RS by learning user preferences based on item popularity or similarity. However, the effectiveness of these methods in mitigating popularity bias and generating novel and unexpected item recommendations remains poorly understood. Recent studies suggest improvement in user preference by incorporating textual opinion provided by the user on an item. Additionally, the trade-off relationship between serendipity's conflicting components, including relevance, novelty, and unexpectedness, warrants further investigation to improve the quality of top-n recommendations. Hence, this research proposes an opinion mining-based approach to learn the users' personalized preferences from the textual reviews and incorporate both rating and reviews preferences to improve the quality of the recommendation list. Next, we design a new Two-Fold Algorithmic (TFA) approach-based objective function for serendipity to mitigate the popularity bias by aggregating uncertainty based on item popularity and item similarity to user preferences. Lastly, a multi-objective evolutionary algorithm-based Serendipity Objective Optimization-based Recommendation Framework(SOORF) is designed to optimize the serendipity's conflicting components. Extensive simulations are conducted over four benchmark datasets. The Mean Absolute Precision(MAP)@n and [email protected] based evaluation findings of SOORF and TFA demonstrate an improvement of at least 8.10% and 58.48%, respectively. The [email protected] and [email protected] based evaluations on different dataset sparsity conditions are observed with an improvement of at least 13.59% and 27.73% over baseline models. The Pareto front shows the models' ability to generate surprising, novel, yet relevant recommendations.

A Deep Autoencoder-Based Hybrid Recommender System

Recommender systems build their suggestions on rating data, given explicitly or implicitly by users on items. These ratings create a huge sparse user-item rating matrix which opens two challenges for researchers on the field. The first challenge is the sparsity of the rating matrix and the second one is the scalability of the data. This article proposes a hybrid recommender system based on deep autoencoder to learn the user interests and reconstruct the missing ratings. Then, SVD++ decomposition is employed, in parallel, to hold information of correlation between different features factors. Additionally, the authors gave a deep analysis of the top-N recommender list from the user's perspective to ensure that the proposed model can be used for practical application. Experiments showed that the proposed model performed better with high-dimensional datasets, and outperformed various hybrid algorithms in terms of RMSE, MAE, and in terms of the final top-N recommendation list.

Research on product recommendation based on matrix factorization models fusing user reviews

Nowadays, recommendation models based on matrix factorization (MF) suffer from the problem of rating sparsity because user-product rating matrix is usually sparse. To address the problem, it is significant to fuse some contextual data or side information on basic MF models. According to this core idea, this paper proposes a modified recommendation model, MFFR (matrix factorization fusing reviews) which recommend products by considering the fusing information on user reviews and user ratings. First, MFFR constructs user-product preference matrix from user reviews by using Latent Dirichlet Allocation (LDA) topic model. Then MFFR predicts ratings and generates personalized top-n recommendation products by using MF model to learn comprehensive latent factors of user-product rating matrix and user-product preference matrix simultaneously. The experimental results of three published datasets demonstrate that our model MFFR can achieve more accurate predicted ratings and hits more correct products of top-n recommendation than the comparative traditional models. MFFR can effectively raise the quality of recommendation, especially in the high level of rating sparsity.

User Demographic Information and Deep Neural Network in Film Recommendation System based on Collaborative Filtering

Research about implementation of deep neural network in recommender system based on collaborative filtering received many attentions recently. One of the major problems in deep neural network based collaborative filtering recommendation system was coldstart problem. Some recent work tried to improve model performance by modifying how the model modelled the interaction between user and item features to generate TOP-N recommendations. This work proposed DNCF (Demographic Neural Collaborative Filtering) model that utilized user demographic information and deep neural network architecture to generate film recommendation system based on collaborative filtering in cold-start problem. NCF model was used as baseline model for model performance comparison. Hit Ratio and Normalized Discounted Cumulative Gain for TOP-10 recommendations were used as evaluation metrics for model performance. Experiment results showed that the proposed DNCF model outperform baseline NCF model by 23,61% in HR@10 and 22,40% in NDCG@10 evaluation metrics. Keywords — collaborative filtering, deep neural network, demographic information, neural collaborative filtering, recommendation system

MulSimNet: A multi-branch sub-interest matching network for personalized recommendation

Personalized recommendation serves as an indispensable functionality in many online services, where the key is to model the user’s preference based on past user-item interactions. Many well-known models learn a latent vector for each user and each item. Nevertheless, a single latent vector might not be able to well represent a user’s multifaceted interests revealed by the wide variety of items consumed by him/her. In this paper, we present a flexible framework named Multi-branch Sub-interest matching Network (MulSimNet) for personalized recommendation. Each user is explicitly represented by multiple sub-interest vectors as well as a general interest vector, and an item is matched against these vectors through different branches, where each branch is realized by appropriate neural networks. The compatibility scores output by multiple sub-interest matching branches are first aggregated via the max operator, which is then fused with the output of the general interest matching branch to estimate the user’s affinity with the item. A comprehensive set of experiments on benchmark datasets demonstrate the effectiveness of the proposed framework in fully capturing the user’s multifaceted interests and yielding better top-N recommendations.

Knowledge distillation meets recommendation: collaborative distillation for top-N recommendation

Knowledge distillation meets recommendation: collaborative distillation for top-N recommendation

Hybrid Sparrow Clustered (HSC) Algorithm for Top-N Recommendation System

Recommendation systems suggest relevant items to a user based on the similarity between users or between items. In a collaborative filtering approach for generating recommendations, there is a symmetry between the users. That is, if user A has similar interests with user B, then an item liked by B can be recommended to A and vice versa. To provide optimal and fast recommendations, a recommender system may generate and keep clusters of existing users/items. In this research work, a hybrid sparrow clustered (HSC) recommender system is developed, and is applied to the MovieLens dataset to demonstrate its effectiveness and efficiency. The proposed method (HSC) is also compared to other methods, and the results are compared. Precision, mean absolute error, recall, and accuracy metrics were used to figure out how well the movie recommender system worked for the HSC collaborative movie recommender system. The results of the experiment on the MovieLens dataset show that the proposed method is quite promising when it comes to scalability, performance, and personalized movie recommendations.

GraphHINGE: Learning Interaction Models of Structured Neighborhood on Heterogeneous Information Network

Heterogeneous information network (HIN) has been widely used to characterize entities of various types and their complex relations. Recent attempts either rely on explicit path reachability to leverage path-based semantic relatedness or graph neighborhood to learn heterogeneous network representations before predictions. These weakly coupled manners overlook the rich interactions among neighbor nodes, which introduces an early summarization issue. In this article, we propose GraphHINGE ( H eterogeneous IN teract and aggre G at E ), which captures and aggregates the interactive patterns between each pair of nodes through their structured neighborhoods. Specifically, we first introduce Neighborhood-based Interaction (NI) module to model the interactive patterns under the same metapaths, and then extend it to Cross Neighborhood-based Interaction (CNI) module to deal with different metapaths. Next, in order to address the complexity issue on large-scale networks, we formulate the interaction modules via a convolutional framework and learn the parameters efficiently with fast Fourier transform. Furthermore, we design a novel neighborhood-based selection (NS) mechanism, a sampling strategy, to filter high-order neighborhood information based on their low-order performance. The extensive experiments on six different types of heterogeneous graphs demonstrate the performance gains by comparing with state-of-the-arts in both click-through rate prediction and top-N recommendation tasks.

A novel top-n recommendation method for multi-criteria collaborative filtering

Most online service providers utilize a recommender system to help their customers’ decision making process by producing referrals. If a customer requests a suggestion for a specific item, the recommender systems produce predictions for it. On the other hand, it is also possible to create top-n lists containing the products that the customer might like the most. Recommender systems’ outcomes depend on individuals’ preferences which can be provided by considering a single criterion or multiple criteria about the services or products. Therefore, there must be methods to produce predictions and top-n lists for single and multiple-criteria datasets. Although the researchers introduced several algorithms on single criterion-based ratings for producing single predictions and top-n lists, there are only methods for producing referrals for a specific item on multi-criteria data. Accordingly, this paper proposes an intuitionistic fuzzy set-based top-n recommender system method with a novel neighborhood formation process for multi-criteria datasets. The proposed method consists of two crucial points: (i) Determining the relational structure between products; (ii) Investigating user tendencies, as well as their distinctive structures and rating distributions. The rating distribution and the relational structure between the products are determined with association rule mining and entropy measure, while the attitudes and tendencies of the users during the evaluation are analyzed with intuitionistic fuzzy sets. We also adopt a single-criterion top-n method to a multi-criteria recommender system, and we employ crisp ratings instead of fuzzy ones to compare the performance of the proposed method. The measurements of serendipity, diversity, and novelty are utilized to show how the experimental results are compelling. When the experiments’ results are examined, it is concluded that our method can generate successful top-n lists.