MACHINE LEARNING STEPS

1. Exploratory Data Analysis

Youtube Link: https://www.youtube.com/playlistlist=PLZoTAELRMXVPQyArDHyQVjQxjj_YmEuO9

STEP 1: EDA { Exploratory Data Analysis }

(i) Check Numerical Features.

(ii) Check Categorial Feature.

(iii) Check Missing Values.

(iV) Check Outliers.

(v) Data Cleaning.

2. Feature Engineering

Youtube Link: https://www.youtube.com/playlist?list=PLZoTAELRMXVPwYGE2PXD3x0bfKnR0cJjN

STEP 1: Handling the Missing Values

• For Numerical Value :

  (i) Mean / Median

  (ii) Random Sample Imputation

  (iii) Capturing NAN Value with a new feature.

  (iv) End of Distribution Imputation.

  (v) Arbitrary Value Imputation (Min / Max) value.

• For Categorical Value :

  (i) Mode (Frequent Category Imputation).

  (ii) Capturing NAN Value with a new feature.

STEP 2: Handling Imbalanced Dataset

(i) Under Sampling ( Best for small dataset ).

(ii) Over Sampling ( Best for large dataset ).

STEP 3: Treating the Outliers

(i) IQR.

STEP 4: Scaling down the Data.

(i) Standardization.

(ii) Normalization.

(iii) Robust Scaler (Scaling To Median And Quantiles).

(iV) Gaussian Transformation.

- Logarithmic Transformation.
- Reciprocal Transformation
- Square Root Transformation.
- Exponential Transformation.
- Box Cox Transformation.

STEP 5: Converting the Categorical Features into Numerical Features.

(i) One Hot Encoding {Nominal}.

(ii) One Hot Encoding with many category in a feature {Nominal}.

(iii) Count or Frequency Encoding {Nominal}.

(iv) Mean Encoding {Nominal}.

(v) Probability Ratio Encoding {Nominal}.

(vi) Ordinal Number Encoding {Ordinal}.

(vii) Target Guided {Ordinal Encoding}.

3. Feature Selection

(i) Variance Threshold ( Unsupervised )

(ii) Correlation

(iii) Chi-Square

(iv) Genetic Algorithm

(v) K Neighbour

(vi) Feature Importance {Extra Tree Classifier}

4. Model Building

1. Supervised Machine Learning

STEP 1: Regression

i. Linear Regression -- Handle Outlier(Y) -- Required Feature Transformation(Y).

ii. Ridge and Lasso Regression -- Handle Outlier(Y) -- Required Feature Transformation(Y).

iii. Decision Tree Regressor -- Handle Outlier(N)

iv. Random Forest Regressor -- Handle Outlier(N) -- Required Feature Transformation(N).

v. XGBoost Regressor -- Handle Outlier(N)

vi. AdaBoost Regressor -- 

vii. ANN --        Required Feature Transformation(Y)

viii. RNN --        Required Feature Transformation(Y)

ix. KNN -- Handle Outlier(Y) -- Required Feature Transformation(Y)

x. SVM -- Handle Outlier(N)

xi. Gradient Boosting -- Handle Outlier(N)

STEP 2: Classification

i. Logistic Regression -- Handle Outlier(Y) -- Required Feature Transformation(Y)

ii. Decision Tree Classifier -- Handle Outlier(N)

iii. Random Forest Classifier -- Handle Outlier(N) -- Required Feature Transformation(N).

iv. XGBoost Classifier -- Handle Outlier(N)

v. AdaBoost Classifier --

vi. ANN --      -- Required Feature Transformation(Y)

vii. CNN --      -- Required Feature Transformation(Y)

viii. Naive Baye's Classifier -- Handle Outlier(N) -- 

ix. KNN -- Handle Outlier(Y) -- Required Feature Transformation(Y)

x. SVM -- Handle Outlier(N)

xi. Gradient Boosting -- Handle Outlier(N)

2. Unsupervised Machine Learning

STEP 1: Clustering

STEP 2: Dimensionality Reduction

i. KMeans -- Handle Outlier(Y) -- Required Feature Transformation(Y)

ii. DBScan -- Handle Outlier(Y)

iii. Hierical Clustering -- Handle Outlier(Y) -- Required Feature Transformation(Y)

iv. KNN Clustering

v. PCA -- Handle Outlier(Y)

vi. LDA

vii. Neural Network -- Handle Outlier(Y)

5. Model Deployment