SP-DDPT: a simple prescriptive-based domain data preprocessing technique to support multilabel-multicriteria learning with expert information

作     者：Mehfooza, M. Pattabiraman, V. 

作者机构：Department of Information Technology Rajalakshmi Engineering College Chennai India School of Computing Sciences and Engineering VIT University Chennai India School of Computing Sciences and Engineering VIT University Chennai India 

出 版 物：《International Journal of Computers and Applications》 (Int J Comput Appl)

年 卷 期：2021年第43卷第4期

页      面：333-339页

核心收录：

学科分类：08[工学] 0835[工学-软件工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Our Sincere thanks to COMAPS  INCOIS  from the Ministry of Earth Sciences  Government of India for providing us dataset for successful implementation of our algorithm 

主　　题：Machine learning 

摘      要：Computing is a consortium of machine learning techniques which works well and provides a sufficient information to make assessment from the chronological historic dataset. In real today a lot of machine analytic techniques available to explore and to investigate the depth knowledge of the dataset but fails to work with multicriteria. There are domains with multicriteria-based decisions like agricultural, spatial, medical, and etc., These domain dataset requires expert intervention as multicriteria to streamline the knowledge presented from the efficient machine learning techniques. The analysis should extend with prescription in order to make productive actions from the learning. This research work is proposed to support the prescription of dataset with a simple automated framework to express the expert percepted value. The simple prescriptive-based domain data preprocessing technique (SP-DDPT) algorithms were defined to support the universal criteria such as ‘less than,’ ‘more than’ and ‘between.’ The proposed framework was applied to the sample Indian coastal dataset and the expert percepted value was preprocessed and prescribed successfully. © 2018 Informa UK Limited, trading as Taylor & Francis Group.