0

I try to predict the result of an personality type test based on how people answered. My sample consists of the answers which range from 1 (strongly disagree) to 7 (strongly agree). Six answers lead to a result. I have the results and answers of 80 people.

An example of my dataset looks like this:

Name AnswerQ1 AnswerQ2 AnswerQ3 AnswerQ4 AnswerQ5 AnswerQ6 Result
Patrick 2 4 7 1 3 4 76%
... ... ... ... ... ... ... ...
Stefanie 1 7 7 6 5 6 93%

I guess I could use multiple interpolation or curve fitting. Does someone know matching python functions for that?

yemy
  • 129
  • This is known as a regression problem. Perhaps that's what you might mean by "multiple interpolation"? – whuber Apr 16 '22 at 20:24
  • Your problem statement is not clear. Is "Result" a function of how the subjects answered questions Q1-Q6? How do you measure Result? – dipetkov Apr 16 '22 at 21:03
  • I am not sure if it is a regression problem. Regression predicts on given X values one target y. I want to do it the other way around. Given y how to predict X. In my case X are the values which represent the answer to an question (1 - strongly disagree & 7 - strongly agree). @whuber – yemy Apr 16 '22 at 21:47
  • Thank you! I added some more information and hope it's clear now. Indeed the "Result" is a function based on the subjects answered questions. @dipetkov – yemy Apr 16 '22 at 21:53
  • Your table indicates "result" is a single number and "six" clearly indicates you have multiple answers, while your question asks how "to predict the result ... based on how people answered." It is difficult to see this as being anything other than regressing the result (a single response variable) against the answers (six explanatory variables), which is a multiple regression problem. If you still disagree, then please edit your post to show explicitly how this is not multiple regression. – whuber Apr 16 '22 at 22:08
  • It's actually even less clear. You cannot predict X from y, because for any one person you don't know their y before they have answered X. – dipetkov Apr 16 '22 at 22:11
  • 1
    You are right. I had a understanding problem. I just apply multiple regression on my existing data. Thank you. @whuber – yemy Apr 17 '22 at 02:07

0 Answers0