Description

Artificial Intelligence (UCS-411)
Lab Assignment-5
1 Create a dataset (.csv file) having following features- experience of the candidate, written score, interview score and salary. Based on three input features, HR decide the salary of the selected candidates. Using this data, KNN model build a for HR department that can help them decide salaries of the candidates. Print all the
possible model evaluation parameters.
Also predict the salaries for the following candidates, by executing the model (for different values of K):
(a) 5 Yrs experience, 8 written test score, 10 interview score
(b) 8 Yrs experience, 7 written test score, 6 interview score
2 Create a dataset (.csv file) having following features- Graduations percentage, experience of the candidate, written score, interview score and selection. Selection feature is binary in nature and contains the status of the candidate. Also store at
least 25 records in this dataset.
Using this data, build a Bayesian learning model for HR department that can help them to decide whether the candidate will be selected or not. Take 80% data as training data and remaining a testing data randomly. Using the built model, predict the status for the following unseen data:
(a) 90 %, 5 Yrs experience, 8 written test score, 10 interview score (b) 75%, 8 Yrs experience, 7 written test score, 6 interview score
Also calculate the possible classification metrics for the above cases and save these values in the .CSV file.
3 For the IRIS dataset, design a decision tree classifier. Take different percentage of training data and then observe effect on the accuracy and other evaluation parameters. Also note the effect of decision tree parameters (like max depth,
min_sample_spit etc.) on the performance of the model.
Note: Take criterion as entropy.

4 Compare the performance of KNN, Bayesian Classifier and Decision Tree model on the diabetes dataset. For this comparison, you can take different possible
parameters of particular model. Save outputs in a .csv file.