Data Preparation

Data preparation is the most important step in Machine Learning.
Almost 70% of the time is spent on data cleaning.

Step 1: Data Collection

Data sources can be:

• Excel file
  

• Database
  

• API
  

• CSV file
  

Example (Student Dataset):

Step 2: Data Cleaning

• Handle missing values
  

• Remove duplicates
  

• Fix incorrect data
  

Example:

• Blank attendance → fill with average
  

• Duplicate rows → remove

Step 3: Feature Selection

Not all columns are useful for the model.

Example:

• Student ID → Not useful
  

• Study Hours → Useful

Step 4: Feature Engineering

Creating new meaningful features from existing data.

Example:

• Convert Attendance % into categories
  

• Calculate Total Score

Step 5: Data Encoding

Machine Learning models cannot understand text.
Convert text into numbers.

Example:

Convert to:

• Pass = 1
  

• Fail = 0

Step 6: Data Splitting

Divide dataset into two parts:

• Training Data (70–80%)
  

• Testing Data (20–30%)
  

Example:

If dataset has 1000 rows:

• 800 → Training
  

• 200 → Testing

Model Training

Now we train the machine using data.

Step 1: Select Algorithm

Choose algorithm based on problem type:

• Regression → Linear Regression
  

• Classification → Logistic Regression
  

• Clustering → K-Means

Step 2: Train Model

The model learns patterns from training data.

Example:
The machine learns:
“More study hours → Higher chance of passing”

Step 3: Model Testing

Use testing data to check whether the model predicts correctly or not.

Step 4: Evaluate Model

Regression Metrics:

• MAE (Mean Absolute Error)
  

• MSE (Mean Squared Error)
  

• R² Score
  

Classification Metrics:

• Accuracy
  

• Precision
  

• Recall
  

• Confusion Matrix