Python and Deep Learning

Lab Assignment-2

TEAM ID :8

*VIDEO 🔗 : https://drive.google.com/open?id=1tYUE3xBI1vcLOAXXxmB6bfatGTo93RWs

NAME 1: Nimmagadda Sudheer

ID: 17

NAME 2: Uppalapati Adisekhar

ID: 21

NAME 3: Mandava Mahesh

ID: 15

Introduction

• This lab assignment is about using classification, clustering, regression and nlp(nltk) on different datasets.
• we observe accuracy for classification, clustering, and regression methods.

Objective

• Our main objective in this assignment is to use a dataset and split it into train and test apart. And use specific classification/clustering/regression method to train the dataset to fit it to the particular corresponding model.
• Finally we measure the accuracy for each of the method.

Approaches

• Given in the lab assignment for classification we have used algorithms like Naive Baye's, SVM, KNN on a chosen dataset and reported the accuracy.
• For clustering we used KMeans on a chosen dataset and visualized the clusters.
• For the nlp question we used nltk and performed different operations like Tokenization, Lemmatization, Trigrams on the given input file.
• For regression we performed multiple linear regression on a chosen dataset and reported RMSE, R^2.

Workflow

• Team member-1: completed question 1, half part in 3rd question and video.
• Team member-2: completed question 2 and other half part in 3rd question.
• Team member-3: completed question 4 and documentation.

Datasets

• http://archive.ics.uci.edu/ml/machine-learning-databases/postoperative-patient-data/post-operative.names

• https://archive.ics.uci.edu/ml/machine-learning-databases/00476/

• http://archive.ics.uci.edu/ml/datasets/Servo

Evaluation and Discussion

QUESTION-1

• We used one hot encoding for categorical features and mean technique for handling null Values.
• SVM classifier got more accuracy than other two classifiers for post operative patient data.

QUESTION-2

• We used elbow method to find optimal number of clusters and evaluated the silhouette score.
• we visualized the clusters using matplot lib

QUESTION-3

• Initially we read data from file and tokenized the text into words and applied lemmatization on each word.
• we also extracted top 10 trigrams based on their count and displayed the sentences with the most repeated trigrams.

QUESTION-4

• we used multiple regression model and used one hot encoding on the chosen dataset and reported the RMSE and R2

• REFERENCES

1. https://stackoverflow.com/questions/tagged/python

2.https://docs.python.org/3/

                                                     THE END