Скачать или смотреть ML-Statistics : Parametric & Non Parametric methods for Correlation

Statistics basics classes in Bangalore by 6benches.in
----------------------------------------------
*Parametric methods(covariance & Pearsonr) and Non-parametric(spearman rank correlation) methods covered in this lecture
a) Covariance(from numpy import cov) ranges from - infinity to + infinity. Which means, we cannot interpret, what is good covariance and what is bad covariance. To overcome this, we use, i)Pearson r linear correlation coefficient(from scipy.stats.stats import pearsonr) and ii)Spearman - non-linear correlation coefficient.
b) Label encoding the categorial data using importing labelEncoder from sklearn.preprocessing and by fit-transform method
c)searborn.lmplot("col1","col2",data=dataset) plots regression model across data
c) Seaborn heatmap to visualize the impact of one column against other column in dataset seaborn.heatmap(dataset.corr())