Cold-start Problem Research Articles

Integrated sentiment analysis with BERT for enhanced hybrid recommendation systems

Recommendation systems play a crucial role in assisting users by providing personalized suggestions based on their preferences. However, these systems often face challenges such as data sparsity, cold start user issues, and overlooking the critical aspect of user sentiment. To address these limitations, this study integrates sentiment analysis into recommendation systems using a dataset from Yelp, focusing on two domains: restaurants and hotels. We conduct preprocessing tasks to facilitate sentiment analysis and leverage the power of BERT, achieving an accuracy rate of 90% for the restaurant domain and 87% for the hotel domain. We then incorporate the sentiment analysis component into collaborative filtering, utilizing model-based deep matrix factorization, and significantly reduce the root mean square error (RMSE) to 0.1 for the restaurant domain and 0.1186 for the hotel domain. Additionally, we integrate sentiment analysis into content-based recommendation using clustering, resulting in improved recommendations with a higher silhouette score of 0.533 for the restaurant domain and 0.42 for the hotel domain. To further enhance system performance, we propose a novel approach that combines these sentiment-aware components using NMF with DecisionTreeRegressor, achieving an even lower RMSE of 0.02 for the restaurant domain and 0.01 for the hotel domain. This integration of sentiment analysis into the recommendation system demonstrates its effectiveness in improving accuracy and personalization, providing users with more meaningful and relevant recommendations based on their sentiments. However, challenges such as data sparsity, cold start problems for new users, and other limitations remain, warranting further research to mitigate these challenges for a more robust recommendation system.

FELRec: efficient handling of item cold-start with dynamic representation learning in recommender systems

AbstractRecommender systems suffer from the cold-start problem whenever a new user joins the platform or a new item is added to the catalog. To address item cold-start, we propose to replace the embedding layer in sequential recommenders with a dynamic storage that has no learnable weights and can keep an arbitrary number of representations. In this paper, we present , a large embedding network that refines the existing representations of users and items in a recursive manner, as new information becomes available. In contrast to similar approaches, our model represents new users and items without side information and time-consuming finetuning, instead it runs a single forward pass over a sequence of existing representations. During item cold-start, our method outperforms similar method by 29.50–47.45%. Further, our proposed model generalizes well to previously unseen datasets in zero-shot settings. The source code is publicly available at https://github.com/kweimann/FELRec.

A self-supervised assisted label-efficient method for online remaining useful life prediction

Remaining Useful Life (RUL) prediction is crucial for safe and efficient production in industrial equipment. Conventional methods struggle in online environments, facing challenges like data distribution differences across working conditions, model cold start problem under new equipment, and lack of labels in online scenarios. To address these challenges, we propose a novel framework for online RUL prediction that leverages knowledge from self-supervised tasks via metric learning. First, a label-efficient self-supervised learning method incorporating soft triplet loss is proposed to transfer degradation knowledge. Second, we develop a two-stage method to distinguish different degradation stages of the equipment for better alignment of data distribution. Third, we introduce a pseudo-label dynamic self-updating strategy and a sample entropy-based data replay strategy, fine-tuning the offline model with newly acquired data. Our method outperforms state-of-the-art unsupervised approaches, reducing RMSE and MAPE by 16.02% and 8.40%, respectively, and achieves performance comparable to supervised methods.

The deep learning-based physical education course recommendation system under the internet of things

This study aims to propose a deep learning (DL)-based physical education course recommendation system by combining the Internet of Things (IoT) technology and DL, to improve the accuracy and personalization of recommendation. Firstly, IoT devices such as smart bracelets and smart clothing are used to monitor students' physiological data in real-time, and IoT sensors are utilized to sense the environment around students. Secondly, IoT devices capture students' social interactions with their peers, recommending socially oriented courses. Meanwhile, by integrating IoT data with students' academic data, course recommendations are optimized to match students' learning progress and schedule. Finally, Generative Adversarial Network (GAN) models, especially the improved Regularization Penalty Conditional Feature Generative Adversarial Network (RP-CFGAN) model, deal with data sparsity and cold start problems. The experimental results show that this model performs well in TopN evaluation and is markedly enhanced compared with traditional models. This study denotes that integrating IoT technology and GAN models can more accurately understand student needs and provide personalized recommendations. Although the model performs well, there is still room for improvement, such as exploring more regularization techniques, protecting user privacy, and extending the system to diverse platforms and scenarios.

E-commerce and Mobile Application Development in the Sports Industry

In order to achieve user recommendations that best match their current contextual needs, the author proposes a mobile service QoS (Quality of Service) hybrid recommendation model based on sports user situational awareness. Cluster the users and service items covered by mobile users based on their location contextual information according to the classification principle of autonomous systems, forming a collaborative filtering and recommendation mechanism for mobile user service items; In response to the cold start problem of new users and new projects in traditional QoS recommendation methods, prediction and recommendation of missing QoS attribute preference values are based on User based and Item based CF; In response to the problem of difficult determination of mixed recommendation weights caused by massive data and uneven distribution of service QoS attribute values in mobile network environments, MapReduce based ant colony neural network weight training is used for CF mixed recommendation. Experimental results have shown that the Hadoop sports industry has improved the operational efficiency of algorithms and reduced the time for global QoS preference prediction; And by comparing the operation results of the 10% and 100% sub datasets, it can also be seen that the acceleration ratio of the algorithm will continuously improve with the increase of data volume, thereby improving the operational efficiency of the recommendation algorithm. It has been proven that the MapReduce based ant colony neural network weight training method significantly reduces the global computation time of the algorithm and improves the operational efficiency of the recommendation system.

Fake It Till You Make It—A Statistical Assessment of the Proportion of Fake Reviews in Closed Reputation Systems

ABSTRACT Fake reviews have become an increasingly relevant issue in electronic commerce. Regulators and platform operators have begun to take action against fake reviews through regulation and litigation. While there has emerged an ample body of research on fake reviews in open reputation systems (i.e. where everyone can write reviews), little is known about the frequency and nature of fake reviews in closed reputation systems (i.e. where the completion of a transaction is a necessary precondition for reviewing). This article assesses the frequency of fake reviews in this particular domain of the platform economy. To do so, we draw on data from Airbnb and investigate systematic differences between first and second rating score distributions, estimating the proportion of potentially fake five-star ratings using KL-divergence. In doing so, we advance the understanding of fake reviews in closed reputation systems, offering theoretical insights into review dynamics and practical strategies for detecting fraudulent behavior using score distributions and user attributes. This shows that early-on reputation is crucial for complementors to overcome the cold-start problem, and platform operators will have to think of ways to allow them to do so (e.g. reputation imports) in order not to jeopardize the overall credibility of their reputation systems.

Addressing cold start problems in new store locations with transfer learning in spatial GNNs

The cold start problem poses a significant challenge for retailers opening new store locations, primarily due to the lack of historical sales data necessary for accurate demand forecasting and effective inventory management. This paper explores the application of transfer learning within spatial Graph Neural Networks (GNNs) as a solution to this issue. By leveraging existing data from established stores that share similar characteristics, our proposed methodology enhances the forecasting accuracy and helps mitigate the risks associated with new store openings. We detail the architecture of the spatial GNN model, which captures complex spatial relationships and customer interactions, providing richer insights into demand patterns. Experimental results demonstrate substantial improvements in forecasting performance compared to traditional methods, highlighting the potential of transfer learning to inform strategic decision-making in retail. This research aims to provide actionable insights for retailers seeking to optimize their operations in new markets.

Deep shared learning and attentive domain mapping for cross-domain recommendation

Cross-domain recommendations (CDR) present a viable solution and are increasingly used to address the cold-start problem. Recently, CDR methods are utilizing deep models to generate latent preferences from context vectors or rating matrices and transfer these preferences between domains. However, many of these models focus on learning latent preferences using domain-related information and often disregard preference patterns from the contrary domain. Incorporating the contrary domain preference patterns into deep models can improve the generation of more effective latent representations. Moreover, existing CDR models face challenges in effectively transferring mapped preferences between domains due to the large features disparity between them. In this study, we tackle these problems and present a novel Deep Shared Learning and Attentive Domain Mapping (DSAM) approach for CDR. Specifically, we propose a variant of Long Short-Term Memory (LSTM) called shared learning LSTM, which incorporates the learning of cross-domain preference patterns alongside domain-specific user/item embeddings derived from textual reviews to dynamically generate shared contextual representations in each domain. We further exploit a multi-head self-attentive network to match item-specific knowledge from the source and target domains into different subspaces. We aggregate this learned knowledge to predict rating scores for cold-start users in the target domain. We efficiently optimize this framework in an end-to-end fashion. Experimental results on five real-world datasets demonstrate the effectiveness of our proposed approach against various groups of recommendation models. Additionally, we provide insights to help understand the model architecture and its robustness in handling cold-start users.

A two-stage recommendation optimization algorithm based on item popularity and user features

Financial product recommendation algorithms are mainly product-centered. This article proposes a two-stage recommendation optimization algorithm based on item popularity and user features, named CPCF-TSP, that can make full use of the demographic characteristics of users and mitigate the problem of users being more inclined to choose “hot” financial products. A popularity weight factor is introduced to normalize popularity and modify Pearson's similarity function. The modified Pearson's similarity function is combined with popularity normalization and user features to improve modeling performance. The two-stage recommendation optimization procedure was combined with a collaborative filtering algorithm to improve recommendation precision. CPCF-TSP fully considers user features in building a hybrid recommendation model and solves the problem of user cold-start. It can also mitigate popularity deviations and improve recommendation precision. MovieLens data and Santander Bank client trading data were used in a case study. The results show that the algorithm reduces inaccuracy in the calculation of the weights for recommendation popularity and similarity and is especially suitable for recommending financial products in which user information can be easily collected and the number of users is far greater than the number of products considered.

Session context data integration to address the cold start problem in e-commerce recommender systems

Recommender systems play an important role in identifying and filtering relevant products based on the behaviours of users. Nevertheless, recommender systems suffer from the ‘cold-start’ problem, which occurs when no prior information about a new session or a user is available. Many approaches to solving the cold-start problem have been presented in the literature. However, there is still room for improving the performance of recommender systems in the cold-start stage. In this article, we present a novel method to alleviate the cold-start problem in session-based recommender systems. The purpose of this work is to develop a session similarity-based cold-start session alleviation approach for recommendation systems. The developed method uses previous sessions’ contextual and temporal features to find sessions similar to the newly started one. Our results on three different datasets show that, based on the provided Mean Average Precision and Normalised Discounted Cumulative Gain scores, the Session Similarity-based Framework consistently outperforms baseline models in terms of recommendation relevance and ranking quality across three used datasets. Our approach can be used to address the challenges associated with cold start sessions where no previously interacted items are present.

New conditional probability-based collaborative filtering for the binary market basket data with the high dimensional cold-start problem

Recommendation systems are crucial for customer relationship management in marketing. There exist a lot of collaborative filtering techniques for recommender systems. The Pearson correlation-based approaches have been utilized for collaborative filtering using the binary market basket data with the high-dimensional cold-start problem. As an alternative of the Pearson correlation-based approaches, we propose a new conditional probability-based collaborative filtering based on Bayes theorem for the binary market basket data with the high-dimensional cold-start problem. The proposed conditional probability-based collaborative filtering outperforms the existing approaches based on the experimental results.

Graph contrast learning for recommendation based on relational graph convolutional neural network

Current knowledge graph-based recommendation methods heavily rely on high-quality knowledge graphs, often falling short in effectively addressing issues such as the cold start problem and heterogeneous noise in user interactions. This leads to biases in user interest and popularity. To overcome these challenges, this paper introduces a novel recommendation approach termed Knowledge-enhanced Perceptive Graph Attention with Graph Contrastive Learning (KPA-GCL), which leverages relational graph convolutional neural networks. The proposed method optimizes the triplet embedding representation of entity-item interactions based on relationships between adjacent entities in a heterogeneous graph. Subsequently, a graph convolutional neural network is employed for enhanced aggregation. Similarity scores from a contrastive view serve as the selection criterion for high-quality embedded representations, facilitating the extraction of refined knowledge subgraphs. Multiple adaptive contrast-loss optimization functions are introduced by combining Bayesian Personalized Ranking (BPR) and hard negative sampling techniques. Comparative experiments are conducted with ten popular existing methods using real public datasets. Results indicate that the KPA-GCL method outperforms compared methods in all datasets based on Recall, NDCG, Precision, and Hit-ratio measures. Furthermore, in terms of mitigating cold start and noise, the KPA-GCL method surpasses other ten methods. This validates the reasonability and effectiveness of KPA-GCL in real-world datasets.

Crossing the chasm: Understanding users’ motivational differences based on stages of online community

Drawing on chasm theory and perceived value theory, this study investigated user participation behavior differences at different stages of online communities to understand two stage-related threats: the cold-start problem (i.e., creating conditions for the establishment of an initial mass of content) and the chasm problem (i.e., user motivational change as an online community evolves). Analysis of survey data from 657 TikTok users revealed changes in user participation motivation that can be associated with the online community's development, providing evidence for the existence of a chasm in the online community context. This research has implications for both theorists and practitioners looking to build sustainable online communities.

Adaptive Knowledge Contrastive Learning with Dynamic Attention for Recommender Systems

Knowledge graphs equipped with graph network networks (GNNs) have led to a successful step forward in alleviating cold start problems in recommender systems. However, the performance highly depends on precious high-quality knowledge graphs and supervised labels. This paper argues that existing knowledge-graph-based recommendation methods still suffer from insufficiently exploiting sparse information and the mismatch between personalized interests and general knowledge. This paper proposes a model named Adaptive Knowledge Contrastive Learning with Dynamic Attention (AKCL-DA) to address the above challenges. Specifically, instead of building contrastive views by randomly discarding information, in this study, an adaptive data augmentation method was designed to leverage sparse information effectively. Furthermore, a personalized dynamic attention network was proposed to capture knowledge-aware personalized behaviors by dynamically adjusting user attention, therefore alleviating the mismatch between personalized behavior and general knowledge. Extensive experiments on Yelp2018, LastFM, and MovieLens datasets show that AKCL-DA achieves a strong performance, improving the NDCG by 4.82%, 13.66%, and 4.41% compared to state-of-the-art models, respectively.

Hybrid attribute-based recommender system for personalized e-learning with emphasis on cold start problem

This article introduces a recommendation system that merges a knowledge-based (attribute-based) approach with collaborative filtering, specifically addressing the challenges of the pure-cold start scenario in personalized e-learning. The system generates learning recommendations by assessing item similarities, utilizing the Rogers-Tanimoto similarity measure for materials and users, and Jaccard's similarity for user comparisons. Unlike traditional collaborative methods relying on prior ratings, this approach depends on attributes. Additionally, user and learning material profiling structures were created to serve as fundamental inputs for the recommendation algorithm. These profiles represent student and material knowledge in a two-dimensional space to facilitate matching. Our processes incorporate user learning styles, preferences, and prior knowledge as metrics for achieving the desired level of personalization. The system produces a list of top recommendations based on predicted ratings. To validate its efficacy, a website resembling our learning platform was developed and tested by users. The primary results demonstrate the system's ability to identify similar users even in a pure cold start condition without existing ratings. Consequently, the system proves its capability in recommending suitable materials, modeling students, and identifying similar user groups. The evaluation results of the proposed system showed a good level of satisfaction by the testimonials, quantified by a score of 82% for the recommended materials (16% higher than exiting cold-start systems), and an average score of 90% in terms of satisfaction about the generated student profiles. As they proved the capability of the framework in recommending suitable materials, and its capability in modeling students, finding similar groups of users.

Implementation and Evaluation of a Hybrid Recommendation System for the Real Estate Market

Introduction: The real estate market has been transformed by digital technologies, especially Industry 4.0, which has made property searching and evaluation more efficient, improving its accuracy with the use of advanced algorithms. Traditional methods have been replaced by online platforms using modern machine learning (ML) algorithms, leading to the need for personalized recommendation systems to improve user experiences. Methodology: This study designed and implemented a hybrid recommendation system that combines collaborative and content-based filtering techniques. The development process involved four phases: literature review, technology selection, prototype implementation, and system deployment. Findings: The proposed hybrid model effectively addressed challenges such as data sparsity and the cold start problem, improving recommendation accuracy. In the evaluation, users indicated high satisfaction with the system’s ability to offer personalized property recommendations. Conclusion: Thus, hybrid recommendation systems can significantly improve the property search experience by offering personalized recommendations. However, further research into the applicability of the system in different domains remains a need for further exploration.

RGMeta: Enhancing Cold-Start Recommendations with a Residual Graph Meta-Embedding Model

Traditional recommendation models grapple with challenges such as the scarcity of similar user or item references and data sparsity, rendering the cold-start problem particularly formidable. Meta-learning has emerged as a promising avenue to address these issues, particularly in solving the item cold-start problem by generating meta-embeddings for new items as their initial ID embeddings. This approach has shown notable success in enhancing the accuracy of click-through rate predictions. However, prevalent meta-embedding models often focus solely on the attribute features of the item, neglecting crucial user information associated with it during the generation of initial ID embeddings for new items. This oversight hinders the exploitation of valuable user-related information to enhance the quality and accuracy of the initial ID embedding. To tackle this limitation, we introduce the residual graph meta-embedding model (RGMeta). RGMeta adopts a comprehensive approach by considering both the attribute features and target users of both old and new items. Through the integration of residual connections, the model effectively combines the representation information of the old neighbor items with the intrinsic features of the new item, resulting in an improved initial ID embedding generation. Experimental results demonstrate that RGMeta significantly enhances the performance of the cold-start phase, showcasing its effectiveness in overcoming challenges associated with sparse data and limited reference points. Our proposed model presents a promising step forward in leveraging both item attributes and user-related information to address cold-start problems in recommendation systems.

Exploring RoBERTa model for cross-domain suggestion detection in online reviews

Detecting suggestions in online review requires contextual understanding of review text, which is an important real-world application of natural language processing. Given the disparate text domains found in product reviews, a common strategy involves fine-tuning bidirectional encoder representations from transformers (BERT) models using reviews from various domains. However, there hasn't been an empirical examination of how BERT models behave across different domains in tasks related to detecting suggestion sentences from online reviews. In this study, we explore BERT models for suggestion classification that have been fine-tuned using single-domain and cross-domain Amazon review datasets. Our results indicate that while single-domain models achieved slightly better performance within their respective domains compared to cross-domain models, the latter outperformed single-domain models when evaluated on cross-domain data. This was also observed for single-domain data not used for fine-tuning the single-domain model and on average across all tests. Although fine-tuning single-domain models can lead to minor accuracy improvements, employing multi-domain models that perform well across domains can help in cold start problems and reduce annotation costs.

Adaptive Preference Measurement with Unstructured Data

Many products are most meaningfully described using unstructured data such as text or images. Unstructured data are also common in e-commerce, in which products are often described by photos and text but not with standardized sets of attributes. Whereas much is known about how to efficiently measure consumer preferences when products can be meaningfully described by structured attributes, there is scant research on doing the same for unstructured data. This paper introduces a real-time, adaptive survey design framework for measuring preferences over unstructured data, leveraging Bayesian optimization. By adaptively choosing items to display based on uncertainty around a nonparametric utility model, the proposed method maximizes information gain per question, enabling quick estimation of individual-level preferences. The approach operates on embeddings of the unstructured data, thereby eliminating the requirement for manual coding of product attributes. We apply the method to measuring preferences over clothing and highlight its potential for both the general task of marketing research and the specific task of designing customer onboarding surveys to mitigate the cold-start recommendation problem. We also develop methods for interpreting the nonparametric utility functions, which allow us to reconstruct consumer valuations of discrete attributes, even for attributes that were not considered or available a priori. This paper was accepted by Duncan Simester, marketing. Fundings: Funding for this project was provided by Analytics at Wharton, the Wharton Behavioral Lab, and the Wharton Dean’s Fund. The author also thanks the Govil Family for financial support. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2023.03775 .

Bayesian Model Selection for Addressing Cold-Start Problems in Partitioned Time Series Prediction

Bayesian Model Selection for Addressing Cold-Start Problems in Partitioned Time Series Prediction