Data Mining Attribute Selection Approach for Drought Modelling : A Case Study for Greater Horn of Africa

Getachew B Demisse,Yared Bayissa,Tsegaye Tadesse

doi:10.5121/ijdkp.2017.7401

Abstract

The objectives of this paper were to 1) develop an empirical method for selecting relevant attributes for modelling drought, and 2) select the most relevant attribute for drought modelling and predictions in the Greater Horn of Africa (GHA). Twenty four attributes from different domain areas were used for this experimental analysis. Two attribute selection algorithms were used for the current study: Principal Component Analysis (PCA) and correlation-based attribute selection (CAS). Using the PCA and CAS algorithms, the 24 attributes were ranked by their merit value. Accordingly, 15 attributes were selected for modelling drought in GHA. The average merit values for the selected attributes ranged from 0.5 to 0.9. Future research may evaluate the developed methodology using relevant classification techniques and quantify the actual information gain from the developed approach.

Highlights

Attribute selection is the process of identifying relevant information and removing as much of the irrelevant and redundant information as possible [1]
The remaining attributes were obtained from the United States Geological Survey (USGS) [17], European Space Agency (ESA) [18], International Soil Reference and Information Centre (ISRIC)–World Soil Information [19], Climate Hazards Group Infrared Precipitation with Stations (CHIRPS) [20], EROS Moderate Resolution Imaging Spectroradiometer [21], National Aeronautics and Space Administration (NASA) EARTHDAT [22], and The Nature Conservancy's (TNC) GIS data portal [23]
The middle of the growing season was selected with the assumption that there would be strong correlation between independent attributes and dependent attributes (SSG), which may help us in selecting the relevant attribute and discarding the irrelevant attributes

Summary

Introduction

Attribute selection is the process of identifying relevant information and removing as much of the irrelevant and redundant information as possible [1]. Attribute selection is the process of selecting relevant drought variables for constructing drought prediction models in space-time dimensions. The attribute selection approach here is focused on identifying and selecting the most relevant drought descriptor variables from different sources and removing the irrelevant attributes without loss of information. One popular categorization has coined the terms filter and wrapper to describe the nature of the metric used to evaluate the worth of attributes [4]. Wrappers evaluate attributes by using accuracy estimates provided by the actual target learning algorithm. On the other hand, use general characteristics of the data to evaluate attributes and operate independently of any learning algorithm [1]

Objectives

Methods

Results

Conclusion