$1.$

Compare and contrast supervised, unsupervised, and semi$-$supervised learning in terms of:

a$.$ Input Data: Describe the type of data $($labeled vs$.$ unlabeled$)$ used in each learning paradigm.

b$.$ Objective: Explain the primary goal of each type of learning.

c$.$ Algorithms: Name at least two algorithms typically associated with each type of learning.

d$.$ Discuss one scenario where semi$-$supervised learning may be more advantageous than using purely supervised or unsupervised learning.

$2.$

a$.$ Explain the concepts of oversampling and undersampling in machine learning. Why are they necessary when dealing with imbalanced datasets?

b$.$ Describe two techniques for oversampling and explain how they work:

o Random Oversampling

o SMOTE $($Synthetic Minority Over$-$sampling Technique$)$

c$.$ Describe undersampling and explain how it works.

d$.$ Advantages and Disadvantages:

o Compare the benefits and drawbacks of using oversampling vs$.$ undersampling. In what scenarios would one approach be preferable to the other?

$3.$

a$.$ Define Agglomerative and Divisive Hierarchical Clustering:

o What are the key steps involved in each method?

b$.$ Dendrograms:

o Explain how dendrograms are used to represent the results of hierarchical clustering.

c$.$ Linkage Methods $($Agglomerative$-$specific$)$:

o Name and describe at least two linkage criteria $($e$.$g$.,$ single$-$linkage, complete$-$linkage$).$ How do these affect the clustering outcome?