Implemented probability estimates of each word using the formula:
Counts of gram sequence in corpus/ Counts of context window sequence in corpus

Improved the I/O logic a bit aswell, the debug statements are there but commented.