Finding contribution/weight of features in output from a series of data

I have a dataset which consists of the following features : a1,a2,.....aN.... sum of all the features together gives a constant output say 100. For eample the dataset may look like this.

a1 a2 a3 a4 a5 a6 a7 O/p

500 100 20 700 0 0 10 100

20 25 5 700 200 300 400 100

0 800 0 0 0 0 0 100

Here the features a1, a2 etc are not identical in nature.I need to determine what values or range of values of the features correspond to what O/p value. For example for simplicity, in the 3rd Row 800 unit of a2 contribute to 100 in the o/p. But as you see the real data would not be this simple. It would be rather like row 1 and row 2. What could be the possible approach to solve this type of problem?Any suggestion would be highly appreciated.

Topic linear-programming feature-engineering regression statistics clustering

Category Data Science

1
answered at

Not sure if I'm understanding it correctly or not.

If you're looking for a way to evaluate the respective predictive power of each features to the O/p value, Weight of Evidence and Information Value might be what could help you.

But if you're trying to find a method or formula to predict O/p value using your features, perhaps it belongs to regression problem and you might want to have a check on regression models.

About

Geeks Mental is a community that publishes articles and tutorials about Web, Android, Data Science, new techniques and Linux security.