Improving Collaborative Filtering Algorithms

International Journal for Modern Trends in Science and Technology, 2021

Collaborative Filtering Using Data Mining and Analysis

With the recent trend of expressing opinions on the social media platforms like Twitter, Blogs, Reviews etc., a large amount of data is available for the analysis in the form of opinion mining. This analysis plays pivotal role in providing recommendation for ecommerce products, services and social networks, forecasting market movements and competition among businesses, etc. The authors present a literature review about the different techniques and applications of this field. The primary techniques can be classified into Data Mining methods, Natural Language Processing (NLP) and Machine learning algorithms. A classification framework is designed to depict the three levels of opinion mining –document level, Sentence Level and Aspect Level along with the methods involved in it. A system can be recommended on the basis of content based and collaborative filtering

International Journal of Electrical and Computer Engineering (IJECE), 2024

Online learning presents a major challenge for learners, namely the diversification of courses and information overload. In response to this issue, recommender systems are widely used. Nowadays, social networks have become a global platform where individuals share a multitude of information. For instance, Twitter is a social network where users exchange messages and interact with various communities. These interactions on social networks have created a new dimension in the field of online learning. In this article, we propose a novel approach that combines sentiment analysis of learners’ reviews on social networks with collaborative filtering methods to provide more personalized and relevant course recommendations. To achieve this, we explored different models to analyze the sentiments of tweets related to online courses. Additionally, we used collaborative filtering based on k-nearest neighbors (KNN). Our results demonstrate that integrating sentiment analysis provides more relevant recommendations. This has also been shown based on the calculation of root mean square error (RMSE) compared to a traditional approach. In this study, we demonstrated that by leveraging this information from social networks like Twitter, online learning platforms can enhance the effectiveness of their course recommendations, tailoring them to each individual learner’s needs.

2018 Fourth International Conference on Information Retrieval and Knowledge Management (CAMP), 2018

The rush to purchase the latest products sometimes prevents people from thinking things through completely. Consequently, recommender services are increasingly emerging. By looking at industry trends, interviewing dozens of leading industry stakeholders, and using publicly available information, it is important to filter out the most relevant information for consumer electronics before purchasing their items. This paper presents an electronic product recommender system based on contextual information from sentiment analysis. The recommendation algorithms mostly rely on users' rating to make prediction of items. Such ratings are usually insufficient and very limited. We present a contextual information sentiment based model for recommender system by making use of user comments and preferences to provide a recommendation. The purpose of this approach is to avoid term ambiguity which is so called domain sensitivity problem in recommendation. The proposed contextual information sentiment-based model illustrates better performance by using results of RMSE and MAE measurements as compared to the conventional collaborative filtering approach in electronic product recommendation.

The Social networking site is increasingly used as a channel for reaching end users.Personalized Recommender system can work on participatory media content and enhance CMC (computer mediated communication) ultimately providing the user with the finest items of interest. It collects data implicitly as well as explicitly and takes into con sideration user activity, preferences, and ratings to evaluate weights for calculation of trust, social intimacy, popularity and semantic scores. The accumulation of these scores generates the final recommendation score and based on it a recommendation list is generated for each user .Several important theories in this regard have proven to be viable and some not so feasible. Thus comparative study of some recommendation systems can throw light on the problems faced and suggest solutions in this regard

International Journal of Research, 2018

In recent years, shopping online is becoming moreand more popular. When its needed to decide whether to purchase a product or not on-line, the opinions of others become important. Product or service retrieval from the huge data becomes very trouble, so users review and ratings are considered as a major point for recommendation and filtering. But, the user’s reviews are very high and this need a lot of time to take decision from the reviews. So, there is a need to summarize the reviews with finding sentiment analysis. This paper presents a survey on various Sentiment Analysis and mining techniques with different applications. From the comparative study, the new and optimal method can be detected and used for web app recommendation with Rating Prediction from huge set of text reviews.

With the introduction of Web 2.0; The users’ generated content like Reviews, Feedbacks, comments, Web Chats, Votes (Likes | Dislikes), Ratings (Stars Ratings) have grown exponentially over time and provided great opportunity for Research Scholars, Organizations, Businesses to mine this useful information and make use of it for variety of novel work like Recommendations. As the time passed, the information overloading problem arrives. There is lot of users’ generated data collected by Organizations and Businesses such as Reviews; How to extract useful information from these reviews and make a perfect recommendation is crucial. Traditional Recommender Systems (RS) considering a number of factors, such as product category, Stars Ratings, Location, user purchase history and other social factors. In this paper we have implemented the Recommender System as proposed by Lei et. al. [5]. The dataset was taken from yelp.com. Model Training is done by Latent Dirichlet Algorithm(LDA) along with Sentimental Dictionaries and Score computation methods as proposed by Lei et. al.[5]. The whole work has been implemented on MatLab (2016a) and experimental results were also analyzed.

systems are a subclass of information filtering system that seek to predict the 'rating' or 'preference' that user would give to an item. Recommender systems have become extremely common in recent years, and are applied in a variety of applications. Collaborative Filtering is one of the most successful approaches to building recommender system, which uses the known preferences of a group of users to make recommendations or predictions of the unknown preferences for other users. The most popular ones are probably movies, music, news, books, research articles, search queries, social tags, and products in general. However, there are also recommender systems for experts, jokes, restaurants, financial services, life insurance, and Twitter followers. The Matrix Factorization methods have been proposed for social recommendation due to their efficiency in dealing with large data sets. This paper provides an overview of recommender systems that include Collaborative, Content and Matrix factorization methods.

—In recent years, we have witnessed a flourish of review websites. It presents a great opportunity to share our viewpoints for various products we purchase. However, we face an information overloading problem. How to mine valuable information from reviews to understand a user's preferences and make an accurate recommendation is crucial. Traditional recommender systems (RS) consider some factors, such as user's purchase records, product category, and geographic location. In this work, we propose a sentiment-based rating prediction method (RPS) to improve prediction accuracy in recommender systems. Firstly, we propose a social user sentimental measurement approach and calculate each user's sentiment on items/products. Secondly, we not only consider a user's own sentimental attributes but also take interpersonal sentimental influence into consideration. Then, we consider product reputation, which can be inferred by the sentimental distributions of a user set that reflect customers' comprehensive evaluation. At last, we fuse three factors—user sentiment similarity, interpersonal sentimental influence, and item's reputation similarity—into our recommender system to make an accurate rating prediction. We conduct a performance evaluation of the three sentimental factors on a real-world dataset collected from Yelp. Our experimental results show the sentiment can well characterize user preferences, which helps to improve the recommendation performance.

The rapid development of the web applications has created numerous opportunities for the people to freely voice their opinions. The explosion of social media and other web applications such as micro-blogging, e-commerce sites, news portals etc. has resulted in large quantities of user generated content in the form of comments, reviews, feedback, recommendations and ratings. How to analyze and summarize the views expressed in such large opinioned text is a new and fast growing field for research. Sentiment analysis of the user generated content can be very useful for the business organizations, individuals and the Government. In the past years sentiment analysis and opinion mining has emerged as one of the popular techniques for information retrieval and web data analysis. This paper presents a survey on sentiment analysis and the related techniques. It also discusses the application areas and challenges for sentiment analysis with insight into the past researches.