yash2mehta / ML_HMM

50.007 Machine Learning Project 2023

Dependencies needed to install

It is strongly recommended to establish a pip environment for executing all the files. The code was generated using python 3.11.4 and any python version 3.7+ should be consistent with executing our code.

We have listed down the imports for the packages we are using in our project code itself. Here's a list of the pacakges for your reference:

1. Pandas
2. NumPy
3. Regular expression

These packages can be installed via the general command:

MacOS/Linux

pip3 install pandas
pip3 install numpy
pip3 install regex

Windows

pip install pandas
pip install numpy
pip install regex

Instructions to run our code

Please follow the below instructions for setting our code:

Utilize git bash (Windows) or powershell or Mac/Linux:

Step 1: git clone https://github.com/sherinksaji/MLProject.git Step 2: cd MLProject Step 3: git checkout main

Please follow the below instructions for running our code as well as evalResult:

Note: If you are running Windows, use python instead of python3 for each command specified below (unless you have aliased it on your system)

Part 1:

For ES dataset:

python part1.py
python  evalResult.py  Data/ES/dev.out  Data/ES/dev.p1.out

For RU dataset:

python part1.py
python  evalResult.py  Data/RU/dev.out  Data/RU/dev.p1.out

Part 2:

For ES dataset:

python part2.py
python  evalResult.py  Data/ES/dev.out  Data/ES/dev.p2.out

For RU dataset:

python part2.py
python  evalResult.py  Data/RU/dev.out  Data/RU/dev.p2.out

Part 3:

For Evaluating ES 2nd-best k:

python evalResult.py Data/ES/dev.out Data/ES/dev.p3.2nd.out

For Evaluating ES 8th-best k::

python evalResult.py Data/ES/dev.out Data/ES/dev.p3.8th.out

For Evaluating RU 2nd-best k::

python evalResult.py Data/RU/dev.out Data/RU/dev.p3.2nd.out

For Evaluating RU 8th-best k::

python evalResult.py Data/RU/dev.out Data/RU/dev.p3.8th.out

Part 4:

In part4, we have two main files you can test on:

1. dev.in
2. test.in

You can choose whichever file to test on by modifying the flag variable (acts as a toggle) in the following code in part4.py:

# set to dev for analyzing dev.in dataset
# set to test for analyzing test.in dataset
flag = 'test'

Similarly, you can choose whichever dataset you would like to train and test for by modifying the lang variable (acts as a toggle) in the followign code in part4.py:

#set to ES/RU for whichever dataset you wish to analyze
lang = 'ES'

Before running the EvalScript, you must run part4.py without debugging.

For ES dataset:

python part4.py
python  evalResult.py  Data/ES/dev.out  Data/ES/dev.p4.out

For RU dataset:

python part4.py
python  evalResult.py  Data/RU/dev.out  Data/RU/dev.p4.out

Running test data for Part 4:

For ES dataset:

python part4.py
python  evalResult.py  Data/ES/dev.out  Test/ES/test.p4.out

For RU dataset:

python part4.py
python  evalResult.py  Data/RU/dev.out  Test/RU/test.p4.out

Team Members

• Mehta Yash Piyush: 1006516
• Sherin Karuvallil Saji: 1005228
• Tang Heng: 1006102
• Venkatakrishnan Logganesh: 1006050