Analisis Statistika Terhadap Penyebab Angin Kencang dan Puting Beliung di Daerah Istimewa Yogyakarta Tahun 2011-2014

Authors

  • Dwi Shinta Marselina Jurusan Statistika Fakultas MIPA Universitas Islam Indonesia
  • Edy Widodo Jurusan Statistika Fakultas MIPA Universitas Islam Indonesia

Keywords:

Strong Winds and Typhoon, Weather Factors and Altitude, Characteristics of Weather Conditions, Logistic Regression Analysis, Predictive Accuracy

Abstract

The purpose of this research are (1) to comply with a researcher suggestion previously for included altitude variable and give better analysis method than previously, (2) comprehend the characteristics of the weather conditions the day before or at the time occurrence of strong winds and typhoon, (3) analyze the factors that most inluence signiicantly the occurrence of strong winds and typhoon, (4) and to know the logistic regression model that has the highest prediction accuracy in explaining prediction occurs whether or not a potential disaster. The variables were used : the occurrence of strong winds and typhoon which was grouped by classiication of clouds, weather factors (air temperature (°C), rainfall (mm), humidity (%), air pressure (milibar), wind direction (°) and wind speed (knot)), as well as altitude. There is an altitude variable, therefore the logistic regression analysis was divided by physiographic components that arrange DIY. The results of this research indicate that there is equality of each factor of the weather conditions the day before or at the time of occurrence strong winds and typhoon. Obtainable the most inluential factor to the occurrence of the disaster, for irst physiographic are temperature, humidity, wind speed, wind direction, and rainfall with 83,3% prediction accuracy. Second physiographic are temperature, wind speed, wind direction, rainfall, and altitude with 72,3% prediction accuracy. Third physiographic are temperature, wind speed, wind direction, and altitude with 73,4% prediction accuracy. Fourth physiographic is wind speed with 79,2% prediction accuracy.

Published

2015-11-15

Issue

Section

Artikel