Rana Dattagupta

A New Approach to Fast Near-Optimal Channel Assignment in Cellular Mobile Networks

This paper presents a novel method for solving channel assignment problems (CAPs) in hexagonal cellular networks with nonhomogeneous demands in a 2-band buffering system (where channel interference does not extend beyond two cells). The CAP with nonhomogeneous demand is first partitioned into a sequence of smaller subproblems, each of which has a homogeneous demand from a subset of the nodes of the original network. Solution to such a subproblem constitutes an assignment phase, where multiple homogeneous demands are assigned to the nodes corresponding to the subproblem, satisfying all the frequency separation constraints. The whole assignment process for the original network consists of a succession of multiple homogeneous assignments for all the subproblems. Based on this concept, we present a polynomial time approximation algorithm for solving the CAP for cellular networks having nonhomogeneous demands. Our proposed assignment algorithm, when executed on well-known benchmark instances, comes up with an assignment which is always within about 6 percent more than the optimal bandwidth, but requires a very small execution time (less than 5 millisecond on a HPxw8400 workstation). The proposed algorithm is very much suitable for real-life situations, where fast channel assignment is of primary importance, tolerating, however, a marginal deviation (6 percent) from the optimal bandwidth.

Fuzzy clustering with biological knowledge for gene selection

This paper presents an application of Fuzzy Clustering of Large Applications based on Randomized Search (FCLARANS) for attribute clustering and dimensionality reduction in gene expression data. Domain knowledge based on gene ontology and differential gene expressions are employed in the process. The use of domain knowledge helps in the automated selection of biologically meaningful partitions. Gene ontology (GO) study helps in detecting biologically enriched and statistically significant clusters. Fold-change is measured to select the differentially expressed genes as the representatives of these clusters. Tools like Eisen plot and cluster profiles of these clusters help establish their coherence. Important representative features (or genes) are extracted from each enriched gene partition to form the reduced gene space. While the reduced gene set forms a biologically meaningful attribute space, it simultaneously leads to a decrease in computational burden. External validation of the reduced subspace, using various well-known classifiers, establishes the effectiveness of the proposed methodology on four sets of publicly available microarray gene expression data.

A Product Recommendation System Using Vector Space Model and Association Rule

This paper presents an alternative product recommendation system for Business-to-customer e-commerce purposes. The system recommends the products to a new user. It depends on the purchase pattern of previous users whose purchase pattern are close to that of new user. The system is based on vector space model to find out the closest user profile among the profiles of all users in database. It also implements Association rule mining based recommendation system, taking into consideration the order of purchase, in recommending more than one product. To make the association rule memory-efficient, cellular automata is used.

Designing an ontology based domain specific web search engine for commonly used products using RDF

In this 21st Century, globalization has become a common knack for all. In most used ways, globalization is made by creating websites. As a result, information volume of internet has increased exponentially [1]. Therefore, it is increasingly becoming difficult to retrieve specific information from the internet. Search engine plays a big role to retrieve information from internet [2][3]. Effective search result for a user given search string always increase search engine efficiency. Most of the search engines are available in the market that produced search results in two parts, i.e. one is highlighted part and the other one is main part. Most of the cases in the main part of their search results produced by appending title tag, URL and meta-tag information or first few words from the Web-page content. While searching of any information for few commonly used products like book, mobile handset, medicine, jewelry, etc., Web-searcher wants to know some basic attributes based on the product. Now, if we display these products basic attributes information in the search result instead of displaying meta-tag or first few words from the Web-page content, then no need to visit all the URLs displayed in the search result. As a result all the Web-searcher saves their time and Web-page download cost. In this paper, we have proposed a design and development mechanism of Ontology based domain specific Web-search engine for commonly used products using Resource Description Framework (RDF).

Mining the web with hierarchical crawlers – a resource sharing based crawling approach

An important component of any web search engine is its crawler, which is also known as robot or spider. An efficient set of crawlers make any search engine more powerful, apart from its other measures of performance, such as its ranking algorithm, storage mechanism, indexing techniques, etc. In this paper, we have proposed an extended technique for crawling over the World Wide Web (WWW) on behalf of a search engine. This is an approach with multiple crawlers working in parallel combined with the mechanism of focused crawling (Chakrabarti et al., 1999a, 2002; Mukhopadhyay et al., 2006). In this approach, the total structure of any website is divided into several number of levels based on the hyperlink-structure for downloading web pages from that website (Chakrabarti et al., 1999b; Mukhopadhyay and Singh, 2004). The number of crawlers of each level is not fixed, rather dynamic in this context. It is determined at execution time on demand basis using threaded program based on the number of hyperlinks of a specific web page. This paper also proposes a focused hierarchical crawling technique, where crawlers are created dynamically at runtime for different domains to crawl the web pages with the essence of resource sharing.

Introducing Session Relevancy Inspection in Web Page

In this paper, we propose a new technique for checking the relevancy of sessions created by visitors on a web-page for measuring the web-page ranking, since session of a web-page is considered as an important parameter for web-page ranking calculation in a search engine. It is assumed that session on a web-page depends on the relevancy of the web-page contents with respect to the requirement. A longer session on a web-page may not yield high relevancy of the web-page, hence a threshold value (THV) is considered for individual web-page based on the contents to avoid the probable noise. The threshold value (THV) is calculated by Keyword Matching Index (Kindex) and Data Transfer Speed of the client-server. The Kindex is measured by implementing fuzzy logic on Pattern Matching of requirement and web-page contents. Field Matching information is fetched through hierarchical database.

Web-Page Indexing Based on the Prioritized Ontology Terms

In recent years, globalization has become one of the most basic and popular human trends. To globalize information, people always publish their documents in the Internet. As a result, the volume of information in Internet becomes huge and it is still growing at an alarming rate. To handle such huge volume of information, Web-searcher uses search engines. However, one of the most practical issues in this area is to design new efficient search engine that retrieves specific information from those pool of information. Therefore, such kind of problem has got an important attention in today’s human life. However, several web researchers are involved to design efficient search engine by optimizing their algorithms, identifying important parameters, etc. Among them, Web-page indexing has been identified as a crucial parameter. Nevertheless, there are several approaches that have been proposed to index Web-pages by the Web researchers.

Fuzzy Based Multi-Agent System Offering Cost Effective Corporate Envi- ronment

We propose a general framework of a hierarchical structure, consisting of several levels of activities, for typical software related corporate hierarchy using multi-agent system. This work identifies the functionality of each level. Each and every level is considered as agent who is described further using fuzzy analysis. Our framework consists of six levels within which first five levels are considered as agents followed by the human interaction at the lowermost level. These agents interact with each other to produce a desired result for the client based on autonomous decisions which are decided through fuzzy reasoning with the help of predefined databases. A layered architecture has been proposed in this paper for showing a corporate office hierarchy in a cost effective manner. In general, management employees of the corporate sys- tem draw a huge amount of money for their activities. Our ultimate aim is to reduce the cost of the existing corporate sys- tem by observing and controlling the behavioral characteristics of each level of hierarchy by replacing typical manual op- erations with agents (1,2). We have presented a case study or practical engineering example along with the description of each agent.

A Voltage-Controlled Free-Running Multivibrator Using a Pair of Regenerative Switches

A voltage-controlled free-running multivibrator is described. In the new circuit use is made of two bistable regenerative switches connected by a single timing capacitor to realize an astable multivibrator. The circuit provides rectangular output waveforms of very short rise and fall times over a wide range. Low output impedance of the regenerative switches permits driving large resistive and capacitive loads. Among other features of the circuit are the facility of logarithmic or linear variations of frequency by a modulating control voltage, its sure-starting nature, and its integrability.

A New Approach to Design a Domain Specific Web Search Crawler Using Multilevel Domain Classifier

Nowadays information published in the internet has become a common knack for all. As a result volume of information has become huge. To handle that huge volume information, Web researchers are introduced various types of search engines. Efficiently Web-page crawling and resource repository building mechanisms are an important part of a search engine. Currently, Web researchers are already introduced various types of Web search crawler mechanism for the various search engines. In this paper, we have introduced a new design and development mechanism of domain-specific Web search crawler, which uses multilevel domain classifiers and crawls multiple domain related Web-pages, uses parallel crawling, etc. Two domain classifiers used to identify domain-specific Web-pages. These two domain classifiers are used one after the other, i.e., two levels. That’s why we are calling this Web search crawler is a multilevel domain-specific Web search crawler.