Multi-View- Deep neural Network-for-Text-Classification

This code is my implementation of the paper: "Multi View network for text classification"

Software: python 3.5 ¶

• The file glove.6B.50d.txt (or other dimensions) Should be downloaded and put in the same folder where the ipython file is located.
• Hardware used for simple implementation: -HP Computer -OS: Ubuntu 16.04 LTS -RAM: 8Gig, Intel® HD Graphics 520 (Skylake GT2) -Processor: Intel® CoreTM i5-6200U CPU @ 2.30GHz × 4

---

-Used Sharcnet Canada cluster for high computing power -Used NVIDIA TESLA K80 GPU - 32G RAM, (Total memory used: 11.25GiB)

Requirements (Python 3.5 packages):

Glove, pandas, numpy, tensorflow, sklearn, nltk

and regular expression package: re

The accuracy of the validation set is now around 80% :

The model can be improved by trying:

• different optimizers
• different learning rates
• diferent regularization (dropout percentage)
• data Augmentation

---