Reference no: EM133769290

Tasks
Submit a completed Jupyter notebook file containing the following steps and related output:

Import all necessary libraries and modules.
Load the dataset into a Pandas DataFrame.
Perform exploratory data analysis (EDA):
Explore and understand the dataset as a whole.
Explore and probe the many rows and columns. Use visualizations, incl. histograms and any other appropriate plots.
Identify and remedy any deficits (e.g., missing, null data) that need to be addressed in order to do any meaningful analysis.
Perform feature analysis, engineering and selection:
Apply encoding as necessary to make the data usable in training a machine learning model.
Apply scaling/normalization/standardization as appropriate. (may be repeated in multiple, iterative experiments)
Perform basic feature correlation analysis (your choice of approach--e.g., Pearson corr. coef. matrix)
If you find any features that appear to yield little or no predictive value, consider discarding them.
Be cautious, though, as simple linear correlation may not reveal the links among features and target.
Perform any other feature selection and elimination methods you feel are appropriate. Must include at least ONE (1) other than feature correlation from 4.3.
If you feel it's appropriate, apply PCA. I would suggest that you wait until training and testing a model before using PCA and then re-training/testing.
Split your data into training and test sets.
Choose a supervised learning algorithm and use it to train a model to predict whether or not a passenger survived based on the features you've selected/engineered.
Calculate and output your training and testing accuracies as well as confusion matrices for training and testing. Improve your model by modifying hyperparameters, if possible (e.g., k in knn).
You may optionally use cross-fold validation for testing and iterative training.
Repeat as many of the steps above as needed to build a more accurate (i.e., testing accuracy) model--varying your application and choice of feature engineering methods and ML algorithms. Continue repeating the process until you're satisfied.