Clustering Techniques for Software Engineering

Abstract

<p>Software industries face a common problem which is the maintenance cost of industrial software systems. There are lots of reasons behind this problem. One of the possible reasons is the high maintenance cost due to lack of knowledge about understanding the software systems that are too large, and complex. Software clustering is an efficient technique to deal with such kind of problems that arise from the sheer size and complexity of large software systems. Day by day the size and complexity of industrial software systems are rapidly increasing. So, it will be a challenging task for managing software systems. Software clustering can be very helpful to understand the larger software system, decompose them into smaller and easy to maintenance. In this paper, we want to give research direction in the area of software clustering in order to develop efficient clustering techniques for software engineering. Besides, we want to describe the most recent clustering techniques and their strength as well as weakness. In addition, we propose genetic algorithm based software modularization clustering method. The result section demonstrated that proposed method can effectively produce good module structure and it outperforms the state of the art methods. </p>

Energy Efficient for Web of Things Based on Reconfigurable Smart Quality Management Scheme

“…A while later, the registered information is transmitted remotely to the base station where the client can screen the water parameters through Zigbee remote correspondence module. In the proposed keen WQM framework [9], [10], a reconfigurable savvy sensor interface gadget that coordinates information gathering, information preparing, and remote transmission is composed. In the proposed shrewd WQM framework, the ultrasonic sensor is observed the water level.…”

Density-based classification with the DENCLUE algorithm

“…The algorithm gives good representations with the noise points. The contour of the clusters is smoother, while the rectangular regions are substituted by a spherical area identified by the radius We approach in our research method two algorithms, the DBSCAN which allows clustering [4] by density, and the DENsity-based CLUstEring (DENCLUE) [5], it was proposed [6] between 1998 and 2000, based on mathematical functions. Although it acquires high complexity with the number of input parameters, it shows acceptable results.…”