Hello, I’ve just finished the rethinking book and wanted to try out modeling a simple regression then package the model to call from an API. I decided to use a current kaggle data set found here: kaggle competitions download -c elo-merchant-category-recommendation I’m trying this one coefficient…

[image] junpenglao: Try something like mu = pm.Normal('mu', alpha + beta_1*x['feature_1_2'].values, sd=1.) Right, shape errors… How about: mu = pm.Normal('mu', alpha + beta_1*x['feature_1_2'].values, sd=1., shape=x['feature_1_2'].values.shape)

Use to rstan and rethinking pacakge. trying to switch to pymc3...getting an error

Questions

jordan.howell2 December 5, 2018, 8:48pm 6

Thank you for the help.

So when I start adding more variables, do I add additional shape arguments for each variable?

Topic		Replies	Views
Type Error on Regression Problem Questions	11	802	January 15, 2019
KeyError Setting Mixture Proportions for Mixture Model Questions	26	1503	May 13, 2019
Error in basic model due to trigonometric functions using math library Questions	10	2625	March 8, 2019
Value error of pm.model Questions	8	619	May 28, 2019
ValueError: design matrix must be real-valued floating point Questions	10	1476	August 1, 2018

Use to rstan and rethinking pacakge. trying to switch to pymc3...getting an error

Related topics