A

Random Forest model gives 95% accuracy on training data but only 78% on test

data.

1. What issue is the model likely

   facing?

2. Suggest two ways to improve

   generalization performance.

Suppose you are tasked to segment customer types in a telecom dataset (no label

available).

1. Which clustering algorithm is

   better suited: DBSCAN or Fuzzy C-Means? Why?

2. How would you evaluate the

   performance of your clustering?

You are given a dataset with

non-linearly separable classes.

1. Which classifier would you choose

   between Logistic Regression, Decision Tree, and SVM? Justify your choice.

    

2. If using SVM, which kernel function

   might be most appropriate and why?

What

does the parameter `C` control in SVM?

The ε in SVR represents

One

key difference between

In

Random Forest, how is the final classification decided?

In

SVM, support vectors are

Decision

Trees use which of the following for classification tasks?

The

two key parameters in DBSCAN are: