Forecasting of Statistical Crisp Weather in India by Exploration of Different Machine Learning Techniques

The winter season starts suddenly in the mid of the Dew and ends in Spring. If we will analyze in a better way, the cold starts in the mid of October and continue up to mid of February month. Generally, some the states like Jammu, Kashmir, Uttar Pradesh, Uttarakhand, Punjab etcetera the people are suffering from a heavy cold. At that time the temperature is as much as decreased so that the old generation people cannot tolerate it. Sometimes also the meteorological department is very busy predicting temperature because in India other natural disasters like floods, droughts, cyclones etcetera are also coming suddenly. So in this investigation work, we have designed an automatic model for crisp weather forecasting which may be considerate to the people, so that they can take precautions earlier. For this work, we have worked for Odisha state for the investigation purpose which uses barometric parameters as well as the machine learning techniques like Support Vector Machine along with Random Forest for our investigation work. The dataset that we have used is IMD which is used for both training and testing. In this model, we have used two prognoses which are based on five and seven days respectively. The contrast result babbles 5 days produce better classification accuracy than 7 days' prediction.