Literature Review Of The Recommendation Systems Designing
In this essay a detailed literature survey and work done in recent years is elaborated. This essay closely surveys the procedure that current researchers are adopting in recommendation systems. Jiyi WU, Lingdi PING, Han WANG, Zhijie LIN, Qifei ZHANG (2008) give introduction of Mobile Ecommerce Personalized Recommender System concept. Jiyi WU et. al put forward the architecture of MEC-PRS. Algorithm of neighbor-based collaborative filtering and item rating based collaborative filtering are analyzed and compared emphatically. Jiyi WU et. al found that item rating prediction based collaborative filtering recommendation algorithm can improve the recommend quality of PRS in performance test, and collaborative filtering recommendation algorithm based on item rating prediction provides better recommendation results than traditional collaborative filtering algorithms.
SongJie Gong, HongWuYe, Heng Song (2009) proposed and algorithm which employs memory-based CF to fill the vacant ratings of the user-item matrix. Then, it uses the item based CF as model-based to form the nearest neighbors of every item. At last, it produces prediction of the target user to the target item at real time. The collaborative filtering recommendation method combining memory-based CF and model-based CF can provide better recommendation than traditional collaborative filtering.
SongJie GONG (2009) presents case-based reasoning technology to fill the vacant ratings of the user-item matrix. And then, it produces prediction of the target user to the target item using item-based collaborative filtering. The recommendation algorithm combining the case-based reasoning and item-based collaborative filtering can alleviate the sparsity issue and can produce more accuracy recommendation than the traditional recommender systems.
Sarika Mittal, Jothi Swarubini Vindhiya Varman, Gloria Chatzopoulouy, Magdalini Eirinaki, and Neoklis Polyzotisz (2010) presents QueRIE, a recommender system that supports interactive database exploration. This system aims at assisting non-expert users of scientific databases by generating personalized query recommendations. Drawing inspiration from Web recommender systems, QueRIE tracks the querying behavior of each user and identifies potentially “interesting” parts of the database related to the corresponding data analysis task by locating those database parts that were accessed by similar users in the past. It then generates and recommends the queries that cover those parts to the user.
Mustansar Ali Ghazanfar and Adam Prugel-Bennett (2010) propose that there are three main types of recommender systems: collaborative filtering, content-based filtering, and demographic recommender systems. Collaborative filtering recommender systems recommend items by taking into account the taste (in terms of preferences of items) of users, under the assumption that users will be interested in items that users similar to them have rated highly. Content-based filtering recommender systems recommend items based on the textual information of an item, under the assumption that users will like similar items to the ones they liked before. Demographic recommender systems categorize users or items based on their personal attribute and make recommendation based on demographic categorizations. These systems suffer from scalability, data sparsity, and cold-start problems resulting in poor quality recommendations and reduced coverage.
Mustansar Ali proposes a unique cascading hybrid recommendation approach by combining the rating, feature, and demographic information about items. Mustansar Ali et. al empirically show that our approach outperforms the state of the art recommender system algorithms, and eliminates recorded problems with recommender systems.
Namita Mittal, RichiNayak, MC Govil, KC Jain (2010) propose a framework based on, application of data partitioning/clustering algorithm on ratings dataset followed by collaborative filtering for developing a Movie Recommender System. The proposed system reduces the computation time considerably and increases the prediction accuracy.
Ying-Chao Zhang, Chao Chen (2010) proposed an improved collaborative filtering algorithm based on bipartite network, degree of nodes and sort of nodes both have been taken into account. And the only need to calculate the top-Nsimilar neighbors for each target item, which take less reaction time. Based on the MovieLens data set the experimental results demonstrate that the algorithm is better than the standard Pearson and Cosine correlation both in the accuracy and computation time.
Kamal Souali, Abdellatif El Afia, Rdouan Faizi (2011) suggest that recommender systems are considered as one of the basic pillars of e-commerce as they help users to take decisions easily. These systems involve a multitude of technique sranging from hybrid filtering mechanisms to techniques derived from statistics or artificial intelligence. In the present paper, Kamal Souali et. al. put forward an improved recommender system that supports ethics in an automatic way without the user intervention.
Kamal Souali, Abdellatif El Afia, Rdouan Faizi, Raddouane Chiheb (2011) recommender systems are widely used not only in e-commerce but in e-learning as well. They are actually made use of in the latter environment to suggest resources and learning materials to learners and, thus, contribute in improving the quality of both teaching and learning. In that paper, Kamal Souali et. al put forward a new recommendation system that provides learners with the most appropriate answers and clues through a request answer module.
Noraswaliza Abdullah, YueXu, ShlomoGeva (2011) integrate collaborative filtering and search-based techniques for recommending these products. Instead of directly recommending products that the user’s neighbors have an interest in to the active user, the proposed technique, named CFRRobin, uses the products as queries to retrieve other relevant products. Then the returned products from all the queries are merged and ranked by using the Round-Robin method, in order to select the final products to recommend.
Experiments conducted on real e-commerce data show that the proposed approach outperforms the Basic Search (BS) and the standard Collaborative Filtering (CFOriginal) approaches, which are widely applied by the current e-commerce applications. The CFRRobin technique also performs better than the Query Expansion (QE) approach that has been proposed for recommending infrequently purchased products.
ShwetaTyagi, and Kamal K. Bharadwaj (2012) propose a new recommendation scheme that combines case-based reasoning (CBR) with collaborative filtering (CF) and incorporates fuzzy trust model.
The CBR methodology is employed to find the most appropriate cluster that forms neighborhood (nbd) set for the active user. Then BD generation process of CBR, based on user rating vector (URV) and clustering, improves system’s scalability to certain extent. Additionally, the proposed scheme allows users to decide which other user’s opinions they should trust more. In this way, the trustworthy users from the set of neighbors suggested by CBR are filtered by applying a fuzzy trust model. As a consequence, only trustworthy neighbors contribute to the final prediction. Experimental results clearly demonstrate that the proposed recommendation scheme (Trust/CBR/CF) outperforms Pearson CF (PCF) and CBR/CF.
