enerally, people use search engine for one of three things, such as, research, shopping and entertainment [1]. For example, Amazon, Yahoo! Shopping, Google product search etc. handles all of the purchasing, shipping and ordering info. People also look up things like videos, movies, games, social networking sites etc. There are many E-commerce site in Bangladesh, they are increasing by size and number day by day [2]. There is no special search engine to search for product all over these sites. Though the size of information rapidly increasing in this field, so we need an efficient Information Retrieval System to make better use. So, a new Information Retrieval System named as "Bangladesh E-commerce Product Search Engine" is introduced in this project. The proposed system can help to find best relevant data from those sites within a seconds.

The search engine is a software system [3] that compares queries to documents and produces ranked result lists of the documents. Search engine must be able to capture, or crawl, many terabytes of data, and then provide sub second response times to millions of queries submitted every day. This software is a program that search documents for specified keywords given through the E-commerce sites [4] and returns a list of the documents where the keywords were found. Typically, Web search engines work by sending out a spider to fetch as many documents as possible [5]. Another program, called an indexer [6] then reads these documents and creates an index based on the words contained in each document. Each search engine uses a proprietary algorithm to create its indices such that, ideally, only meaningful results are returned for each query. The search engine activities and the core information retrieval issues are shown in Figure -1.

The architecture of e-Search Engine is designed to ensure that a system will satisfy the application requirements of search engine are effectiveness (quality), means it able to (i) effectiveness (quality), that is we want to able to retrieve the most relevant set of documents possible for a query; and (ii) efficiency (speed), that is, we want to process queries from users as quickly as possible [7].

Figure1: Search engine design and the core information retrieval issues [8]. Search engine components support two major functions, such as, the indexing process and the query process [8].The indexing process builds the structures that enable searching. The query process uses those structures and a person's query to produce a ranked list of documents.

The text acquisition component is used to identify and make available the documents that will be searched. Text acquisition will more often require building a collection by crawling or scanning the sources of information. Then the documents to the next component data store, which contains the text and metadata for all the documents. The text transformation components transforms documents into index terms or features and then the documents are stored in index and used for searching. Also, the index creation component takes the output of the text transformation component and creates the indexes or data structures that enable fast searching. This part process user query and find document using index and rank the document for the efficiency of search engine. The proposed search engine involves BM25 ranking algorithm [9] for calculating the rank of a document, shown in Figure -3.The major components of the query processing are user interaction, ranking and evaluation, as shown in Figure -4 [8].

The user interaction component provides the interface between the person doing the searching and the search engine. The main task is to accept user's query and transform it into index terms and then to take the ranked list of documents from the search engine and organize it into the results shown to the user. The ranking component takes the transformed query from the user interaction component and generates a ranked list of documents using scores based or a retrieval model [9].The task of evaluation component is to measure and monitor effectiveness and efficiency. An important part of that is to record and analyze user behavior using log data. Evaluation is primarily an offline activity, but it is a critical part of any search application.

Web crawler, database and the search interface are the major component [10] of a search engine that actually makes search engine to work in the Web. The structure of the Web is shown in Figure -5 which include the crawler, the parser and the downloader [11].Search engines are programs that search documents for specified keywords and returns a list of the document. The web crawler program is to search for the information in the database. Once web crawler finds the pages, the search engine then shows the relevant web pages as a result. A Web crawler is an Internet bot which systematically browses the WWW [12]. A web crawler is designed to follow the links on web pages to discover and download new pages. Web crawlers are also known as ants, automatic indexers, bots, spiders, web robots, and worms. The crawler is adapted to discover and update all documents related to a company's operation.

The design methodology of the proposed system is described by the data flow diagram, as shown in Figure-6.To develop the proposed system, several methods have been implemented in Python programs, such as, Downloader, Parser, Crawl, Tokenize, Build_Index, Query_Transformation, Scoring_Algorithm, etc.

The Downloader(url,retry) method is used to download resource from given URL. The Parser (HTML, ALLOWED_DOMAIN<DEPTH) method parses resource returned by download. The Crawl(allowed_domain, depth) method uses other two methods Parser and Downloader for crawling. The Tokenize (data) method performs text transformation and makes them ready to build index. The Build_index(data)method builds inverted index from transformed data. The Query_transformation(query) method is used to transform user query into tokens so that we can use them with inverted index. The scoring_algorithm(query) transforms user query so that to make them compatible to indexed data. Also a user friendly interface (see Figure -7) has been developed to interact with the proposed that the users can easily access the system. commerce Product Search Engine, so that the proposed system can help to find best relevant data from those sites within a seconds. Though small amount of document and poor text transformation technique search result are not efficient as we expected. The web crawler could not fetch many information correctly, because E-commerce sites use a lots amount of JavaScript, and to get that data the system need to interact with some JavaScript event. This may be improved by using an intelligent web crawler and better text transformation technique. Also spell checking and making query suggestions may be employed further.