Untitled

Abstract

Meteorological data analysis is obtaining the information from raw data. There is vast amount of data available for weather analysis. Market needs timely and accurate data. The collection and datawarehouse of weather data is important because it provides an economic benefit but the local or national economic needs are not as dependent on high data quality as is the weather risk market. The semi arid region of Karnataka namely Madikeri region is considered for data analysis. The relative importance of features is identified for analysis of rainfall data. The adaptive boosting random forest classifier is applied to generate decision rules governing the prediction of rainfall. The data is collected from Indian Meteorological Department (IMD) for span of 12 years from 2004 to 2016. There are 4825 samples considered for the data analysis. The number of features considered for data analysis is 13 for prediction of rainfall. The validation curve and RMSE values justify the results obtained.