IJENS Publisher Indexed In SCOPUS


Title: System for Managing Self-Administered Satisfaction Survey Using Intelligent Agents
Author(s): Blaise Ndangang, Bernabé Batchakui, Claude Tangha, Marvin Te,
Pages: 1-7 Paper ID:170302-8484-IJECS-IJENS Published: April, 2017
Abstract: This article presents a management approach supported by intelligent agents. It is a contribution of artificial intelligence in supporting companies to study the views of their customers. The interest here is focused on the function of optimizing the management process of satisfaction survey questionnaires. This process goes from survey design to the result analysis. The multi-agent system (MAS) developed consists of agents each acting at key stages of the process, allowing an investigator to be interested only content of the survey and result evaluation after administration of the questionnaire to customers. The results are currently being validated in a large satisfaction survey of the Information System Direction (ISD) internal customers of “Les Brasseries du Cameroun et Filiales”.
Keywords: Intelligent agent, investigation, study, survey, satisfaction.
Full Text (.pdf)  International Journals Of Engineering and Sciences | 673 KB
Title: A Framework for Outlier Detection Using Improved Bisecting k-Means Clustering Algorithm
Author(s): K.Swapna, M. S. Prasad Babu
Pages: 8-12 Paper ID:171102-5858-IJECS-IJENS Published: April, 2017
Abstract: The aim of this paper is to design an automatic liver diagnosis system to detect liver diseases early and accurately to help reduce the increasing deaths caused by liver diseases. With this automatic diagnosis system early diagnosis can be done and treatment can be made easy and immediately. Physical data considered for this study is collected from various pathological laboratories from southern India and annotated from expert Gastroenterologists. One of the dataset considered is from Indian liver patient dataset (ILPD) available in the UCI machine repository has 583 records and the physical data collected has 500 records which form a total of 1083 records. Automatic diagnosis tools may reduce burden on doctors. Since common attributes were found in both the datasets considered for the analysis, this paper evaluates the selected outlier algorithms and clustering algorithms using the proposed frame work for clustering liver patient datasets without outliers. These algorithms are evaluated based on four criteria: Accuracy, F-measure, Entropy and Purity. Our interest is to analyze these datasets which would contribute to better understand the system and help us develop an Automatic liver diagnosis system.
Keywords: Liver datasets, Outlier detection, Cluster-based Bisecting k-Means, cluster validation.
Full Text (.pdf)  International Journals Of Engineering and Sciences | 329 KB