saumya-guptaa / ISCO-630E

Soft-Computing

This contains my submissions for the Soft Computing Elective Course at IIITA.

Dependencies

1. Python3
2. Numpy
3. Matplotlib
4. Matlab(Only for Assignment 9)

Assignments

The assignment folders contain the description of the assignment statement along with pyhton codes(and dataset).

1. Using Naive Baysian Classifier: Predict where a given mail is spam or not. Use the data set provided for this purpose. (structured data set).

2. River/Not-River

   1. Using Naive Bayesian classifier predict river non river using Satellite data set of Hooghly river (unstructured data set).
   2. Using PCA: Apply PCA on given river/non-river images.

3. Perform Linear Regression on the given housing dataset with regularization. Also implement LWR and find out what happens when the value of tau is very small.

4. Perform Face Recognition:

   1. Using PCA : Create face dataset using your mobile phone for your face as well as faces of 9 other friends. Create multiple variants (at least 5) of each faces with different view angles.
   2. Using LDA : Create face dataset using your mobile phone for your face as well as faces of 9 other friends.Create multiple variants (at least 5) of each faces with different view angles.

5. Use the microchip dataset . Use 70% of the data for training and 30% for testing.

   1. Use raw data as given, and from there develop a GDA model (without Box-Muller transformation).
   2. For two features, first using Box- Muller transformation create new data set having Gaussian distribution within the range of the given data set and create Gaussian Discriminant Analysis (GDA) model.

6. Supervised Artificial Neural Network

   1. Implement Perceptron training algorithms for AND ,OR, NAND and NOR gates. How you will verify your trained algorithms?
   2. Using two input one output X-NOR data , train a Neural Network using Back Propagation Algorithm.Explain how will you test the network.

7. A Bidirectional Associative Memory is used to store given 4 pairs of patterns of setA and setB. Calculate Weight Matrix and test the level of weight corrections.

8. A Kohennen Network is used to classify 2-D i/p vectors. Training is to be done with randomly generated neurons.

9. Design FIS to diagnose whether a person has Covid'19 or not based on symptoms, using both Mamdani and Sugeno Approaches.