#5: Scikit-learn 3: Preprocessing 3: Scaling a sparse matrix, CSR, CSC format скачать в хорошем качестве

#5: Scikit-learn 3: Preprocessing 3: Scaling a sparse matrix, CSR, CSC format

The video discusses what is a sparse matrix, how it can be compressed to CSR and CSC formats, and scale using Scikit-learn in Python. Timeline (Python 3.8) 00:00 - Welcome 00:14 - Outline of video 01:33 - Sparse vs. Dense Matrix 01:37 - CSR (compressed sparse row matrix) 02:47 - CSC (compressed sparse column matrix) 04:13 - Centering sparse data destroys sparse structure 06:25 - Open Jupyter notebook 07:21 - Create compressed sparse row matrix: .csr_matrix() 07:53 - Create compressed sparse column matrix: .csc_matrix() 08:23 - MaxAbsScaler: fit and transform 09:31 - MaxAbsScaler: fit and transform: using CSR format 10:24 - MaxAbsScaler: fit and transform: using CSC format 11:19 - StandardScaler: fit and transform: 'with_mean=False' 11:53 - * * * ERROR/ TYPO: "scalar" instead of "scaler" because 'scaler' is used in earlier cell and it is printing values of MaxAbsScaler in output instead of StandardScaler * * * 12:24 - StandardScaler: fit and transform: using CSR format 12:48 - StandardScaler: fit and transform: using CSC format 13:00 - Ending notes Correct output for time point 11:53 or 12:05 StandardScalar(): [[0.32349832 0. 0. ] [0. 2.12132034 2.12132034] [2.26448824 0. 0. ]]