Day 26 (19-04-20) : SHALA2020 ASST on Data Science
Completed one assignment of SHALA2020 with begineers testing code of pandas and numpy on movie dataset and train dataset of people who left an org after undergoing changes.
Done all the Data Preprosseing and Visualization on both the datasets.
Day 35 (28-04-20) : Structuring Machine Learning Projects
Complete the course 'Structuring Machine Learning Projects' by Andrew NG under depplearning.ai.
The course consists of 2 Week lectures followed by quiz which tests the approach one needs to go through while implementing the model, studying the overfitting problem, judging variance and bias under training and testing with their solutions
Day 36 (29-04-20) : Decision Trees on Titanic Dataset
Preprocessed the data while dealing with missing values under Age, Cabin and Embarked. Label encoded some of the values then procceded with the model
Implemented a model to predict the Survival of a person using Random Forest Classifier, XG Boost, Gradient Boosting,Decision Tree Classifier and Logistic Regression.
Compared their results,f1 score, percision value and ROC curve for all the three classifiers.
XG Boost works best on the given dataset with 87% test accuracy whereas RFC gives 83% and Gradient Boost gives 83%.
Decision Tree classifier gives 83 and LR gaves 82.
The Ruc of the dataset is 88.25 and XG Boost works pretty close.
Day 37 (30-04-20) : Convolutional Neural Networks: Step by Step
Completed Week 1 of Convolutional Neural Networks by Andrew NG
Worked on implementing Convolution Network networks from scratch which including writing the function of convolution layers, Pooling Layers, padding functions for forward and backward propogation.
Worked on the same application through tensorflow which is one the best framework fo implementing models providing the required libraries and already implemented functions to save time implementing model from scrath.