Implement a simple Bayesian Spam filter and determine if each email in the 'test' folder is spam or ham.

$-$ In the 'train' folder, two files contain $100$ spam and $100$ non$-$spam messages.

$-$ In the 'test' folder, two files contain $20$ spam and $20$ non$-$spam messages.

$-$ We want to classify $40$ emails in the 'test' folder based on the emails in the 'train' folder.

$-$ Thus, probabilies should be calculated from emails in the 'train' folder.

$-$ Please ignore all special characters$($e$.$g$.\text{'}$~$!$@#$$\%^$&

$-$ Use C or C$++.$ No other programming language is allowed. $($ex$.$ python, Java, etc$)$

$[$Procedue for decision.$]$

Per each email from test forder

calculate r$($w$1,......,$ wn$)$ and apply a threshold.

assign a label $($spam or non$-$spam$)$ which is a predicted label.

Calculate the accuracy of your prediction.

$($the number of correctly classified test emails$)$ divided by $40.$

$-$ since we have $40$ test emails $(20$ spam and $20$ non$-$spam$).$

For fun, let's try various 'threshold $($T$)\text{'}$ for decision.

T $=0\mathrm{.}6,0\mathrm{.}7,0\mathrm{.}8,0\mathrm{.}9,0\mathrm{.}95.$

This means that we will have $5$ accuracies.