### Multilayer Binary Classification Problem: Titanic Data

In this assignment, you are expected to create a multilayer neural network model using the Titanic dataset. The model will be developed to predict whether the passengers survived or not. The descriptions of the features and the target variable in the given dataset are given below:

$|$ Feature Name $|$ Description $|$

$|-------------------------|----------------------------------------------------------|$

$|$ Survived $|1=$ alive, $0=$ not alive $|$

$|$ Pclass $|$ Passenger's ticket class $(1=$ First, $2=$ Second, $3=$ Third class$)|$

$|$ Sex $|$ Passenger's gender $(0=$ Male, $1=$ Female$)|$

$|$ Age $|$ Passenger's age $|$

$|$ Family$\_$Size $|$ Number of siblings$/$spouses $+$ parents$/$children on board $|$

$|$ Fare $|$ Passenger's ticket fare $|$

#### Things to do:

$1.**$Data Preprocessing:$**$

$-$ Load the preprocessed titanic$\_$cleaned.csv file.

$2.**$Model Setup:$**$

$-$ Create an artificial neural network model with $5$ input features $($Pclass$,$ Sex, Age, Fare, Family$\_$Size$).$

$-$ Use $64,32$ and $16$ neuron layers as hidden layers and add ReLU activation functions.

$-$ Use sigmoid activation function in output layer.

$3.**$Optimizer Selection and Training:$**$

$-$ Train the model in three different ways using $**$Stochastic Gradient Descent $($SGD$)**,**$Adam$**,$ and $**$RMSprop$**$ optimization algorithms.

$-$ Complete the training on $10,000$ epochs for each optimizer and compare the losses and accuracies.

$4.**$Results Evaluation:$**$

$-$ Compare the results of each optimizer algorithm and determine which algorithm gives better results for this problem.

#### Research Questions:

$1.**$Optimizer Comparison:$**$ Compare the advantages and disadvantages of SGD$,$ Adam and RMSprop algorithms for this problem. Which optimizer converged faster and why?

$2.**$Learning Rate:$**$ How does the accuracy and convergence of the model affect when the learning rate is changed? Compare the results by trying different learning rates.

$3.**$Accuracy Value:$**$ Experiment until the accuracy value of the model is $95\%.$

#### Submission:

$-$ A Jupyter Notebook file with the codes running.

$-$ Graphical comparisons of the accuracy rates and losses obtained for each optimizer.

$-$ A report containing detailed answers to the research questions.