Data Science Online Tutoring

Question 1 Divide the data set to Train, Test Data Sets
Question 2 Use Naive Bayes Classifier for predicting whether the customer was male or female?
Question 3 Use K-Nearest Neighbor for predicting whether the customer was male or female?
Question 4 Compare two models using sensitivity, specificity, ROC and lift chart

Question 1

Divide the dataset to Train and Test Datasets

Question 2

Construct a decision tree and use it for predicting whether the customer is male or female?

Question 3

Construct a random forest and use it for predicting whether the customer is male or female?

Question 4

Compare two models using sensitivity, specificity, ROC and lift chart

Use FamilyIncome , EdYears, Cosmatics

Question 1

Using Hierarchical Clustering how many clusters of customers do you detect?

Question 2

Using K-Means Clustering how many clusters of customers do you detect?

Question 3

Compare and contrast the two Machine Learning models.

Use the train data and Naive Bayes Classifier for predicting whether the customer will buy electronics based on family income and family size.

Use the test data and your model and make predictions regarding whether the customer will buy electronics based on family income and family size.

Question 3a

Use the train data and K-Nearest Neighbor Classifier for predicting whether the customer will buy electronics based on Educational Years and family size.

Question 3b

Use the test data and your model and make predictions regarding whether the customer will buy electronics based on Educational Years and family size.

Compose the confusion matrix of Naive Bayes Classifier and K-Nearest Neighbor decide which model has better accuracy.

Question 4b

Compose the ROC , gain and lift charts of the two models and argue which model is better?

Use k-means clustering, what would be an optimum model that would cluster the buyers based on family income and family size? develop a K-Means model for clustering and Visualize your clusters

Using Hierarchical Clustering what is optimum number of clusters of customers do you detect based on EdYears and FamilySize? develop a Hierarchical clustering model and Visualize your clusters