This Chapter analyzes the recommender systems, their history and its framework in brief. The current generation of filtering techniques in recommendation methods can be broadly classified into the following five categories. Techniques used in these categories are discussed in detail. Data mining algorithms techniques are implemented in recommender systems to filters user data ratings. Area of application of Recommender Systems gives broad idea and such as how it gives impact and why it is used in the e-commerce, Online Social Networks (OSN), and so on. It has shifted the core of Internet applications from devices to users. In this chapter, issues and recent research in recommender system are also discussed.
TopIntroduction
It is the recommender system which is considered one among the most powerful tools in the present digital world. Explanations are usually provided by it to their recommendations so that web users are helped to find its products, people and also their friends who are missing in social communities. In the field of recommender system, there are various methods and approaches which have been implemented. There are two approaches which are most widely used. They are content-based and collaborative approaches. These personalized approaches should be studied so that the best recommendations are provided to the end users.
Originally, we define Recommender systems as ones where ‘‘recommendations are provided by people as inputs, which are then aggregated and directed to appropriate recipients by the system’’. The clear main purpose of the current recommender systems is to guide the user to useful/interesting objects. Due to this, evaluation of recommender systems shows to what extent this goal has been achieved.
Explanations of its recommendations are usually provided by Recommender system so that users are helped better to choose products, activities or even friends. It is the task of recommender systems to turn data on users and their preferences into predictions of possible future likes and interests of the users. When an explanation is received by a user, a recommendation can be accepted more easily since transparency is provided by the system to its recommendations (which follows most of recommender algorithms). The Human Style, Feature Style and Item Style approaches are followed by the most traditional approaches. Even this simple approach can be realized in various ways.
There was an emergence of the first analysis paper on collaborative filtering in the mid-1990s. Since then, Recommender systems have become an important research area basically, recommender systems directly help users to identify content, products or services (such as books, digital products, movies, web sites etc...) with the aggregation and analysis of suggestions from other users, which also means the reviews from a number of authorities and users (Figure 1).
Filtering Technique is said to be the backdrop of every recommender system approach. It gives a clear working nature of this recommender system. This inter disciplinary approach shown here facilitates item ontology and it contains the details of the item for user. The most wanted purchases are watched carefully by the users in order to recommend with clear observation and also to enrich their product promotion. The concentration of item ontology is on the diversified items for the want of the need. The characteristics of each item, along with all recommendation, are given in order to provide a clear vision on the products which are searched. All these data are sent to the recommender systems and they act as filter to the user so that they are able to meet with their needs and purpose. The user modeling data looks up for their likes and interests and they are interpreted with rating in the user rating. The recommender system is the filtering technique. It now filters the data as per the need and interest of the user. They guide users to select the correct content to the correct person. Besides helping to decide the buying products, the recommended item also provides the list of availability to the need of the customer.
TopClassification Method
Our classification framework comprises of recommendation field and data mining techniques. In this research, we segregate the research papers that were analyzed into eight groups of application fields and eight groups of data mining techniques. The general graphical classification framework for recommender systems research papers is shown in Figure 2.
Figure 2. Recommender systems framework