Naive-Bayes-Classifier_with_Normality_Tests

Gaussian Navie Bayes Classifier was applied on IRIS dataset. Different types of normality tests were used to introduce the normality concepts. Implementation of Gaussian Naive Bayes Model is done to demonstrate its working.

First, we need to import the libraries

Pandas, Numpy, Seaborn, Matplotlib, Sklearn, Scipy

Second, need to prepare the data

1. Exploratory Data Analysis

   profile = ProfileReport(df, title='Profile Reports', explorative=True)

2. Standardization : Through StandardScaler

3. Checking Null cells

4. Train/Test Split

Third, Normality Test *check p value according to tests

1. Skewness - Kurtosis Test

   print(stats.kurtosis(X))

   print(stats.skew(X))

2. Shapiro - Wilk Test

   stat, p = shapiro(X[i])

3. Kolmogorov-Smirnov Test

   stat1, p1 = kstest(X[i], 'norm')

Fourth, Applying Gaussian Naive-Bayes Model

Model accuracy was found to be 97.7%

Fifth, Confusion Matrix & Classification Report

1. Confusion Matrix Plot (heatmap)

   sns.heatmap(confusion_mat, annot = True)

2. CLassification Report

Finally, Predict the outcomes