Keras Regression for Deep Neural Networks with RMSE (4.3)

[@sandipprabhu]

Great video series. I tried zscore on all scale variables 'income', 'aspect', 'subscriptions', 'dist_healthy', 'save_rate', 'dist_unhealthy', 'age', 'pop_dense', 'retail_dense', 'crime', and could get excellent results on MSE: 0.002, RMSE: 0.049.

[@nasiksami2351]

Sir you are amazing! Lucky that you are sharing your knowledge with us. Thank you!

[@martinjohnborja1191]

Hi, Jeff. It's been 5 years but I still find this content useful. May I ask what version of tensorflow, pandas, numpy, and sklearn libraries you used to execute the notebook included in the video?. Thank you very much!

[@siddvideos]

Thanks Mr. Jeff, I was making a transition from SciKitLearn to Neural Networks. Your explanation on regression was very helpful. I do have a follow up question, how do you choose number of layers? How to you decide on making that layer dense or not? How are inputs and outputs for each layer are chosen? I apologize in advance if these questions seem really dumb.

[@Perleques]

Great Explanation!You should have more views!

[@zhonghuahe3556]

thanks very much for creating these high-quality videos. one question Mr. Heaton, why don't we do z_score transformation for columns such as 'dist_healthy', 'dist_unhealthy' and 'pop_dense'.

[@mohammadmedhat7379]

Thanks a lot this was so helpful

[@NitinGuptalko]

Hi Jeff, I got a doubt. i see in your video that mean squared error on test data is coming around 30 when you are training the model. however when you calculate score on test data, it is coming 0.45 something. why so much difference, not able to understand that.

[@060584saurav]

Nice video sir.. However I think the error metrices must be calculated after inverse transformation predicted value compared with actual value

[@Diamond_Hanz]

Yeh buddy!