Scripts for the various Artificial Intelligence lab tasks
The task is to perform Part-of-Speech tagging by building a Hidden Markov Model. the build model is dumped into pickle a file which is retrieved during testing
Dataset consists of words on different lines of a sentence which in turn are separeted by a blank line.
After building the model, viterbi algorithm is applied to find the maximum probability tag for every word in the testing datset. The testing datset shoul consist of words that are in the model (smoothing hs not been performed on the model).
The task is to perform Linear and Logistic regression on regression and classification datasets respectively. The model once built is dumped into pickle file.
For linear regression, learing rate is taken to be 0.001 with 200000 iterations.
For Logistic regression , learning rate is taken to be 0.015 with 60000 iterations.
This lab is the application of artificial intelligence in the field of Natural Language Processing. The task is to convert the output from stanford parser(a probabilistic context free grammar,PCFG output) into a dependency tree. Further a naive method has been added for implementation of Paninian Grammar, it describes relation between verb and the objects(nouns/pronouns).
Script is written in python2.7. Specifically it uses anytree module to print the dependency tree. Module can be installed using the command pip install anytree.