Data Analysis project "TITANIC SURVIVAL"
- 2. Import CSV file and start exploring the dataset Check the dimensions of the dataset Call the first rows to understand the data Check the variables and their types Explore the dataset for Missing values and identify their location and their number Subset our data to obtain observations that contain no missing data by replacing the lines with N/A with some specific values Create new variables to help to train our data Exploratory DATA Analysis
- 3. My predictive model is based on some new variables, that I had to create due to make my predictions more accurate. CREATE New variables
- 4. Out-of-bag (OOB) error : is a method of measuring the prediction error of random forests .So the smaller the error, the more accurate my model. Random Forest: Random forest builds multiple decision trees and merges them together to get a more accurate and stable prediction. IMPORTANCE OF VALUES Random Forest has a feature of presenting the important variables. 83.95% Accuracy rf.label <- as.factor(train$Survived) 0:perish 1:survive
- 5. Using 10 fold Cross-validation divide my train.data(=891 obs.) into 10 folds with almost the same length each one Fold1(89 obs.), Fold2(89 obs.),…..,Fold10(90obs.) makeCluster(6,type=“SOCK”) We call these groups 1 to 10 .The analysis is performed 10 times . The first time the analysis is performed, groups 1 to 9 are used to train the algorithm and group 10 is used to test the model. I categorize the data (Clustering) into 6 sockets . Every socket commits CPU and it’s working without waiting at the same time. Performing 10 fold C-V find the cp-accuracy of the model
- 6. Visualization Visualize the data is a powerful machine to understand well your data and find the correlation between them. This plot help me to find the survivng rate based on Pclass and new title. My final rpart model which give me the most important variables of my predictive model. As a result the best accuracy for my model.