Forecasting Weather Using Machine Learning

Forecasting Weather Using Multinomial Logistic Regression, Decision Tree, Naïve Bayes Multinomial, and Support Vector Machine

---

Data set

Our dataset looks like below which we collected from Bangladesh Meteorological Department

We had last 30 years [1988-2017] of weather data.The training and test set is divided into two segments having 70% and 30% data split across the two categories.

---

Parameters:

• Day
• Month
• Year
• Humidity(%)
• Max Temp(in ⁰C)
• Min Temp(in ⁰C)
• Rainfall (in mm)
• Sea Level Pressure (in mb)
• Sunshine (hours)

10. Wind Speed(knot) 11. Cloud (in okta) --- ### Train and Test accuracy of the models tested:

| Model | Training Accuracy (%) | Testing Accuracy (%) | |-------------|------------|------------| | Logistic Regression | 74.2 | 76.9 | | Decision Tree | 76.8 | 74.05 | | Multinomial NB | 54.27 | 54.34 | | SVM | 76.42 | 77.52 |