output1.csv : 4/100/99.7 output2.csv : 2/15/99.8314 output3.csv : 2/15/99.93395 output6.csv : 7 : 99.90
file1.csv: 99.37812 / XGBRegressor(objective ='reg:linear', colsample_bytree = 0.3, learning_rate = 0.1, max_depth = 10, alpha = 10, n_estimators = 10)
Observations:
-
dew_point is highly correlated with visibility_in_miles (ρ = 1). So, one column rejected?
-
Maximum days no rain or snow.( as there are lots of zeroes)
-
Most days are cloudy, windy (as no. of zeroes is low)
-
Temperature value has one 0 Kelvin value
Categorical: Date-Time, weather_description, weather_type, is_holiday
Date-Time: split into Date and Time