479-Project-2

Group members:

• Naoki Atkins [email protected]

• Luke Duggan [email protected]

• Madison Jordan [email protected]

---

KNN Parallelization in Author Classification

In this project, we created a program that will classify an unknown book to an author given a dataset that contains already known books with labels to their authors.

---

Text Preprocessing

run each time a new dataset needs to be generated (if a new book is added).

• input folder (book text files): \books
• output folder (dataset files): \data

First run top.py to find the "bag of words" from each book. After that, you'll need to run set.py so that each "bag of words" can be combined into a single table.

1. python3 top.py
2. python3 set.py

Compiling

mpic++ -fopenmp main.cpp

Run Program

mpirun -n 4 ./a.out

note: n processes should be the number of books in the set. (used in cosine calculation before finding KNN). k is a separate value from n. k is defined within the code at this time (k=3).