Task
Language is extremely expressive. A person can easily discern whether someone is expressing their thoughts coherently just by briefly skimming the language and structure they use. This level of comprehension is my goal.
I want a machine learner to be able to parse text and classify it as helpful or not, to decide whether the content would strike someone as engaging and affective, or mindless, enraging and excessive banter. To be able to do this would grant companies and developers a quick and easy way to pick out the reviews, suggestions, and other forum posts that most likely contain useful feedback out of the hundreds of thousands of posts they receive every so often.
The process I went by in attempting to find a solution employed natural language processing and vector math principles to describe the words and sentences in a review in order to then classify it as helpful or not.
I want a machine learner to be able to parse text and classify it as helpful or not, to decide whether the content would strike someone as engaging and affective, or mindless, enraging and excessive banter. To be able to do this would grant companies and developers a quick and easy way to pick out the reviews, suggestions, and other forum posts that most likely contain useful feedback out of the hundreds of thousands of posts they receive every so often.
The process I went by in attempting to find a solution employed natural language processing and vector math principles to describe the words and sentences in a review in order to then classify it as helpful or not.
Report
project_report.pdf | |
File Size: | 376 kb |
File Type: |
cluster_preprocessing.py | |
File Size: | 4 kb |
File Type: | py |
word_vectorization.py | |
File Size: | 9 kb |
File Type: | py |
arff_file_creation.py | |
File Size: | 9 kb |
File Type: | py |
Contact
Machine Learning, EECS349
McCormick School of Engineering, Northwestern University
McCormick School of Engineering, Northwestern University