International Journal of Computer Applications (0975 – 8887) Volume 94 – No. 18, May 2014 36 A Survey on Techniques for Personalization of Web Search Himani Arya Department of Computer Science MANIT, Bhopal (MP) India Jaytrilok Choudhary Department of Computer Science MANIT, Bhopal (MP) India Deepak Singh Tomar Department of Computer Science MANIT, Bhopal (MP) India ABSTRACT The World Wide Web (WWW) is growing exponentially per year thereby search engine provides the low quality of search results. Thus, the users get difficulty in getting the relevant information from the obtained search results. The quality of web search results depends on the information needs of the user and the searching techniques employed in the web search systems. So, the personalization is a general need in web search now-a-days. This paper includes the review of various approaches towards personalization. The approaches include hybrid profiling, personalized click model, ontology based user profiles and fuzzy theory for personalization. Keywords Personalization, web search, user profile, ontology. 1. INTRODUCTION Internet is a global networking infrastructure that connects computers of many smaller networks. It transports the content to the linked computers all over the world. Traveling of data over the internet occurs in some format known as protocols like SMTP, HTTP etc. The content present on Internet in the form of hypertext that references to the other text or hypertext. This collection of documents around the world forms the web also known as World Wide Web (WWW). The Pages in the hypertext document are known as web pages and its location is specified by a uniform resource locator (URL). Thus, web can be treated as the software which allows the user to use the content available on the internet and write their own content. The web pages on Internet are growing rapidly day-by-day. Thus, an information retrieval tool is needed to find the information form WWW. Web search engine, is an information retrieval tool, needed to look for these web pages. When the user fires a query to any search engine like google or bing, he/she may get even hundreds or more search results related to the query. Generally, theses are divided into a number of search engine results pages (SERPs). From the set of SERPs, user can decide for the link one should try to see if its referenced page contains the desired data. A huge amount of documents is available in web. So, it cannot identify the most important documents for a particular user for a specific query. Due to this reason, the need for personalizing web search arises to find out relevant information for users. Personalization of web search is the process of customizing web search results based on users’ past behavior [1]. Most of the queries submitted to search engines are short [2, 3] and have ambiguity [4, 5]. Every users may have different needs and goals under the same query. Thus the effectiveness of a personalization of web search depends on the query, user and search context [6]. 2. NEED OF PERSONALIZATION Generic Search Engines present the results which are general and not adaptable to individual users. For a particular query fired to the search engine, different results are provided for different users. Search results are organized for every user considering one’s interest, preferences and information needs. The need for personalization arises due to the following facts: firstly, different users have different backgrounds and interests. For the same query, they have different information needs and goals. Secondly, User information needs may change over time. Users may have variety of requirements based on the time and circumstances. For example, a zoologist user may use query ‘‘mouse’’ to find information about computer periphiral when he/she wants to buy a computer mouse and a computer user may submit same query to find the information about the mouse as rodents while watching any animal tv channel. search engines can not to differentiate between such cases. 3. PERSONALIZATION APPROACH When applied to search, personalization would involve the following steps: 1. To collect and represent information about the user in order to understand the user’s interests. 2. Use this information to either filter the results returned from the initial retrieval process, or directly include this information into the search process itself to select personalized results [7]. Web search personalization systems use gathered information about user from profiles, cookies and to conduct and revise the search to maximize the user satisfaction [8]. The user profiles are created which specifies the user’s interests, preferences and information needs to better personalize the search results. There are two ways to generate user profiles- explicit and implicit user profiling. In the explicit approach users create their profiles manually by providing some kind of feedback to a search system. In implicit user profiling, the user profile is created from user’s past behaviour, such as by determining the documents they do select for viewing, the duration of time spent viewing a document or page browsing or scrolling actions [9]. This is being done in the background automatically by the search system. Personalization of web search can be done at either server side or client side. Many problems arises on personalizing the web at server side like server should maintain all the search history for each and every user. It also has to search the history of a particular user when a user submits any ambiguous query. The performance of the server gets down when many users submits the query at the same time. Therefore, most of the techniques employ client side approach as all the search histories and queries are maintained at the client system making the faster way to access the user profile.