Uncategorized

scikit learn – What to do when machine learning regression results in almost the same number for many predictions?



I’m new to machine learning
I have 200 rows dataset 14 daily sensor reading(day 1 to day 14) from daily monitoring, because it’s a rare phenomenon so I only get 200 dataset rows.
Target range(from dataset column) is 100 to 2500 and 600 stdev with normal distribution.
The reggression result always close to 1000(which is the mean/avg & median from the target data).

What I already tried:

  • Check and clean the data(again)
  • using gridsearch for looking the best model parameters
  • try 3 different model(AdaBoost, Gradient Boosting, Randdom forest)
  • The reggression result still close to the mean/avg & median from the target dataset.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *