A Meta Heuristic Classification Algorithm With Multi Objective Based Swarm Intelligence Techniques For Large Structured Information

Abstract

In almost every sector, nature provides as a rich source of newlineinspiration for completing complicated and difficult computational tasks. newlineIn the age of computer, bio-inspired algorithms that replicate natural newlinebehaviour play a crucial role in addressing optimization problems. newlineNumerous research works on optimization, especially in the area of data newlinemining, have been presented in the recent few decades. One of the most newlinecrucial methodologies in data mining is clustering. Because of its lengthy newlineprocessing time, traditional clustering is inefficient. As a result, an newlineoptimization-based approach is used in this thesis for mining the data in newlinean efficient manner. newlineFurthermore, because log files are the sole of data access newlinemethod that stores many events during runtime, they are widely newlineemployed in modern software management activities. Each software newlinesystem event is recorded in the form of log messages, which are made up newlineof a fixed and a variable element. There are various variances in log newlinemessage formats for complex and changing systems, some of which are newlineoften unknown or change on a regular basis. newlinevii newlineAs a result, this research initially utilizes the MapReduce k newlinemean methodologies for clustering the datasets. In addition, the newlineMapRedLGC module is presented, which integrates LGC (Local newlineGravitational Clustering) with MapReduce k mean plays a vital role in newlinesplitting data points into clusters based on Euclidean distances. Then the newlineresearch focuses on using bio-inspired optimization strategies for newlinemachine learning based on data categorization. Next, a hybrid elephant newlineherding opposition methodology is developed, and a comparison analysis newlineis carried out to assess its efficiency. Tests on 21 popular benchmark newlinefunctions are used to evaluate the performance of the proposed method. newlineFinally, a LTD-MO approach is designed which is a unique multiobjective newlineoptimization-based log-file template identification technique. It newlinesolves the challenging optimization problem using a new multi-objective newlinebased on swarm intelligence approach called chicken