Machine Learning Project

Topic : Real Estate Price Prediction

PPT : Presentation

Source Code : Click Here

ML_DUCS

Lab Work :

Lab Assignment - 1 : Apply Decision Tree Classification to the dataset mentioned in the given notebook. Record the accuracy, precision, recall, and f1-score when the data is:

1. Not scaled/ normalized
2. scaled but not oversampled
3. scaled and oversampled

Repeat the same on another dataset from the UCI repository for the binary classification problem.

Lab Assignment - 2 : Implement two algorithms:

1. Find-S
2. Candidate Elimination

---

The following contains Practical Assignment of ML course , Semester - VI , BSc (Hons) Computer Science , University of Delhi

Disclaimer : The following code provided is intended for educational and informational purposes only. It is crucial to use this code responsibly and in compliance with all applicable laws, regulations, and ethical guidelines. The code should not be misused, modified, or repurposed with malicious intent, including but not limited to cheating, hacking etc.

Question 1 Classify the iris dataset using a decision tree classifier. Divide the dataset into training and testing in the ratio 80:20. Use the functions from the sklearn package. Display the final decision tree.

Question 2 Classify the iris dataset using a Bayes classifier. Divide the dataset into training and testing in the ratio 80:20. Use the functions from the sklearn package. Assume the data follows a gaussian distribution. Display the training and testing accuracy, confusion matrix.

Question 3 Classify the iris dataset using the KNN classifier. Divide the dataset into training, validation, and testing in the ratio 70:15:15. Use the functions from the sklearn package. Find the best value for k. Normalize the dataset before applying the model. Display the training, validation, and testing accuracy, confusion matrix.

Question 4 Create a linear regression model using ordinary least squares estimation. Find the best fit line for the dataset ‘salary.csv’ using the above model. Display the training and testing dataset in the scatter plot and draw the best fit line in the same. Also find the MSE and R2 for the testing dataset.

Question 5 Consider the dataset california_housing from sklearn . Find the correlation b/w the different attributes of this dataset. Using the least square estimation method from sklearn, find the best fit line. Also find the error.

Question 6 Consider the dataset ‘Adveristing.csv’. Find the correlation coefficient between the input attributes TV, Radio, Newspaper and Output Attribute Sales. Use least square estimation method to find the line of regression b/w

1. TV and Sales

2. Radio and Sales

3. Newspaper and Sales For all of the above options, also draw a scatter plot and line of regression. Also find the error in each of the above.

Question 7 Consider the dataset ‘Adveristing.csv’. Find the best fit regression line between the input attributes TV, Radio, Newspaper and Output Attribute Sales using gradient descent method. Also find R2 .

Question 8 Use logistic regression to build a model to classify the breast cancer dataset Divide the dataset into training and testing in the ratio 70:30 . Print the confusion matrix, sensitivity, specificity. For each iteration of training, store the training and testing accuracy. Plot a graph showing training and testing accuracy Vs iteration no. Do not use sklearn logistic function.

Question 9 Using logistic regression to build a model to classify the iris dataset. Divide the dataset into training and testing in the ratio 80:20 . Print the confusion matrix, sensitivity and specificity. Alternative

Question 10 Create a linear regression model using the gradient descent method. Create a class to represent the model with the following functions - init, fit and predict. Find the best fit line for the dataset Also find the MSE and R2 for the testing dataset.

Question 11 Consider the dataset wine from sklearn. Using PCA reduce the dimensionality of the dataset to 5. Build a classification model using gaussian naive bayes classifier. Find the training accuracy and test accuracy.

Question 12 Consider the dataset iris. Apply the PCA method to select the best 2 features. Using these features plot the scatter graph. Apply k-means clustering algorithm to cluster the transformed dataset into 3 clusters.

Question 13 Write a program to implement a single layer perceptron model. Train this for solving a AND problem with 3 variables.

Question 14 Consider the dataset iris. Apply hierarchical clustering algorithm to cluster the dataset into 3 clusters.

Question 15 Write a program to implement 2-layered ANN for classifying digits datasets from sklearn. Use 70% data for training the model and check the accuracy of the model on remaining 30% data. Use softmax activation function in the last layer and relu function in the hidden layer.