Engenharia de atributos básica no Keras avaliações
10959 avaliações
Man L. · Revisado há almost 2 years
Elvira S. · Revisado há almost 2 years
Hrutik P. · Revisado há almost 2 years
This lab was pretty bad to be honest. Many of the parts we're expected to fill in were not really covered (e.g., defining a function within a function for scaling the data seems like an over-complicated way of standardizing the data). It also doesn't follow best ML practices. For instance, you train one model, then load the test dataset to make a prediction. Then after feature engineering, you predict on the test dataset again to compare models. This should not be done. Also, you refer to one of the models as 'linear regression', but I don;t understand why? You create a DNN in Keras and use ReLU activation functions, so how is it linear regression?
Robert S. · Revisado há almost 2 years
The jupyter lab did not start
Sonali M. · Revisado há almost 2 years
Wojciech Ś. · Revisado há almost 2 years
Thiago S. · Revisado há almost 2 years
Anthony H. · Revisado há almost 2 years
Very odd prediction for the first model (without feature engineering), house value is predicted below 10000, very far from the solution's 190k. Moreover it seems like the ocean proximity categorical feature is not included in the feature columns in the first part
Etienne V. · Revisado há almost 2 years
Piotr T. · Revisado há almost 2 years
Konrad S. · Revisado há almost 2 years
Amanda K. · Revisado há almost 2 years
Nelson F. · Revisado há almost 2 years
does not work, scikit-learn warning at the beginning
Adrian K. · Revisado há almost 2 years
Kaita F. · Revisado há almost 2 years
Karen L. · Revisado há almost 2 years
Rainer B. · Revisado há almost 2 years
Tulkin Y. · Revisado há almost 2 years
Sandesh T. · Revisado há almost 2 years
Ian C. · Revisado há almost 2 years
Vaibhav N. · Revisado há almost 2 years
Mouad A. · Revisado há almost 2 years
Fabien D. · Revisado há almost 2 years
Akash C. · Revisado há almost 2 years
Yomna M. · Revisado há almost 2 years
Não garantimos que as avaliações publicadas sejam de consumidores que compraram ou usaram os produtos. As avaliações não são verificadas pelo Google.