Implementation of web product search engine (part II)

Abstract

Electronic Commerce has gained popularity in the recent years. It allows consumers to

directly buy goods from a seller over the Internet without an intermediary service at any

time of the day. Customers are attracted to online shopping not only because of the high

level of convenience, but also because of the broader selection and greater access to

information.

A successful web store is not just a good looking website with dynamic technical features. It

is about understanding the customer’s wants and needs, and providing excellent user

experience in assisting customer in quickly seeking out items they are looking for. This leads

to the popularity of a search engine which aims to answer a query within the shortest

possible time with the most relevant results.

Besides the functionality of categorical browsing through hyperlink click access on the E-

Commerce site, the capability of the search engine allows users to search for their desired

products through text format query and image format query. These functions work

individually and may be subjective. An additional useful option for users is the ability to

have a combination of different search types to formulate as a single query.

This project focuses on the addition of various functionalities in the backend Data

Management Server and the Text Server for the web store known as Visual Image Search

Engine (VISE). MySQL database was also modified for better retrieval of records. Techniques

were used to ensure that the increase in components would not affect the performance of

system in returning results in the lowest possible retrieval time.

The system architecture was redesigned to build a system consisting of better role-defined

components with the Data Management Server as the main central server for the front end

to send requests and receives responses from the back end. JSON was used in replacement

of XML to transmit data output from the Text Server as the result from JSON is smaller. The

source codes in DM Server are also rewritten and reorganized to improve code efficiency,

maintainability and feasibility, ensuring that new modules can be easily integrated into the

future project development.