Personalized collaborative filtering recommender system using domain knowledge

Abstract

In the current era of web applications such as e-retail business, the web services focused to provide personalized search systems to the targeted user intents based on the navigation patterns. Intelligent collaborative filtering recommender system tries to recommend the web pages considering the similar patterns of the other users along with the usage knowledge of the current user session. This recommender systems strategy lacks of the domain knowledge in comparing the usage patterns of the other users in serving with recommendations. This paper mainly focused on incorporating the domain knowledge and usage knowledge in personalization as well as in comparing the similar user patterns for recommender systems. This novel strategy builds a model to recommend the web pages that can help the new search scenarios and can improve the likelihood of a user towards the host website. Experimental results shown that the proposed novel strategy yields to gain in performance of the recommender system in terms of the quality of the web page recommendations.

Identification of Astrological belief using Sentimental Analysis by Capturing Opinions from Cross-Domain Individuals

“…SentiSW is another analytical tool that performs Sentiment analysis using entity recognition at the time of sentiment classification [20]. For better analytical results, we can prefer domain ontology with usage knowledge as well as collaborative filtering recommendations [18]. Ms-Excel tools are optimal for our current astrological dataset and the tool is easily downloadable for 32 as well as 64 bit excel and also having compatibility with other operating systems.…”

“…A recommendation system is also required to evaluate facts from captured cross-domain opinions that are in an unstructured input format. In general the evaluation depends on collaboration from distinct recommender systems like personalized web recommendations using ontology [17], collaborative filtering using domain knowledge [18], Intent-aware contextual recommendation system [19] and others.…”

Importance of Domain Knowledge in Web Recommender Systems

“…When a user visits a New Page, Modern recommender system can offer recommendations to the user by mapping the existing web page sequences in the web logs onto the domain term sequences extracted using Domain Ontology [16]. Initially all the web pages are mapped to keywords (Domain Terms) that best represent the information contained in the web page.…”

“…Web page titles are the easiest and most widely used approach to extract domain terms because it is assumed that the web page titles are named as per the content of the web page and they are informative enough. This task is automated using software agents, most widely used of which is protégé [16].…”

“…The domain term sequences are extracted from the web sessions. The domain term sequences once extracted are then mapped onto the Web pages by applying an algorithm that maps the sequences of domain terms to Web pages.This is done page by page [16]…”