This is a predictive model it seeks to predict the production of a set given certain characteristics.
- I made a exploratory data analysis to choose data transformation and feature engineering to feed the predictive model.
-
Train of a model this model predict the production with a error of
+- 803.20 unities
If we consider the dataset average production as 2000 unities, we have a error around of 40% regarding this.
The model has good performance to small productions but when it increases the production the model begins to increase its error, this problem can be solved with a greater amount of data, with a better feature selection or trying with a scaler less susceptible to outliers.