Titanic-Survivability-ML-based-Project

To predict the probability of survival of people who embark on a journey in the Titanic ship. The dataset consists of important features which were to be learnt such as Passenger class, Age, Sugar condition, Physical Injury, etc.The target feature is to predict is the Survival probability whose value is either 1 (can survive) or 0 (cannot survive). Tasks performed: ● Extracted and pre-processed a dataset of passengers list who sailed the Titanic using Python library pandas. ● Implemented Univariate analysis by plotting Histograms with the help of python library matplotlib. ● Implemented Bivariate analysis by plotting Scatter plots with all the considered Feature vectors with the Target vector to identify the relations among the vectors using matplotlib. ● Utilized various supervised learning algorithms such as K-NN Classifier, Logistic Regression, Decision Tree Classifier and Random Forest Classifier to train and evaluate the model. Output : The accuracy score of Test data (30 percent of whole dataset) for four ML algorithms were successfully calculated. Firstly, K-N Neighbours Classifier was used which gave accuracy of around 78 percent for ‘k’ value as 5 for the pre-processed ML model, similarly for Logistic Regressor algorithm the accuracy was about 83 percent. Hence, for Decision Tree Classifier the accuracy score was 90 percent, and for using Random Forest Classifier the accuracy was 81 percent. Hence it can be concluded that the Decision Tree algorithm gave the highest accuracy for predicting the most accurate values for ‘survival’ target vector. This project comes under the category of Supervised Machine Learning Algorithm.