One-Hot Encoding allows encoding categorical variables in numerical ones. This method represents each category of a variable as one column, and a 1 is assigned if the value belongs to the category or 0 otherwise.
Classes, functions, and methods:
df[x].to_dict(oriented='records')- convert x series to dictionaries, oriented by rows.DictVectorizer().fit_transform(x)- Scikit-Learn class for converting x dictionaries into a sparse matrix, and in this way doing the one-hot encoding. It does not affect the numerical variables.DictVectorizer().get_feature_names()- returns the names of the columns in the sparse matrix.
The entire code of this project is available in this jupyter notebook.
|
The notes are written by the community. If you see an error here, please create a PR with a fix. |