Challenge #157: An Expert Challenge

[OllieClarke]

Absolutely would have avoided this question in the exam.

Spoiler

I found the location of the variable importance plot on the community, but then didn't get anywhere calculating the Chi-Sq effect. Had to look at a spoiler to find the nested test tool

[ZenonH]

I spent a good bit of time trying to figure out a way to dynamically select the top 10 predictors, but learned my ability to pick apart R code within Alteryx is not good. 

Spoiler

 

[johnemery]

Definitely a difficult question. It goes to show how tough it can be to know what each predictive tool outputs, not to mention when each tool is appropriate to use.

Spoiler

[mbogusz]

Spoiler

[maheensayeed]

This was a tough one, had to google a number of things to figure it out, but ultimately managed it. This link really helped.
https://community.alteryx.com/t5/Alteryx-Designer-Discussions/Help-Mean-Decrease-in-Gini-for-dummies...

Also hadn't used the nested test tool before, so this was good practice. 

[dsmdavid]

Definitely need to investigate the predictive district. Spent a while moving things from one side to the other before caving in.

Spoiler

[KMiller]

Solution attached. Just thinking this through, as the chi-square result is statistically significant at all thresholds, we are saying the models are significantly similar and the removal of f38 had no statistically significant effect. This is based on the understanding that the null hypothesis for the chi-square test is that there is no association between the results of the two models and the alternative hypothesis (supported by the results) is that there is an association between the two models.

[JamesCameron]

Would never have got this without Google or the reply to this post by @SydneyF. I also had a peak at @pjdit answer to get the logistic rregression tools working.

Spoiler

[atcodedog05]

Great learning !!!!! 🙂 

Spoiler