• Clone the repo git clone <SSH>

• Change to repo's directory cd Job-Recommender

• Requires python 3.7 or more.
• Install virtualenv using pip install virtualenv
• Set up python virtual environment for the project, using virtualenv venv
• Activate the virtualenv.
• If you are using windows then use the command ./venv/scripts/activate.
• If you are using linux systems then use command source venv/bin/activate
• Download the required dependencies used in the project using, pip install -r requirements.txt

Congratulations! You have successfully set up the environment!

• The Datasets/ directory contains all the datasets downloaded for the project. It mainly consists of two job datasets one for IT and other for Non-IT.
• Data preprocessing and natural language processing is done in code/data_cleaning.py.
• Saved cleaned data to a new directory Cleaned_Datasets/ to save time and reduce redundancy during runtime.
• Job recommender is built in Code/model.py

nltk library used. The following pipeline has been executed.

• remove all non alphabets regex = [^a-zA-Z],
• remove whitespaces
• convert case to lowercase
• tokenize words
• remove stopwords
• stemming The result saved to Cleaned_Dataset/ directory.

• First vectorize the text using tfidf vectorizer.
• Then use cosine similarity to find the similarity of text and the one with highest score is the recommended job.

• First copy the contents of your resume or a resume whose job you want to predict and paste the copied contents in predictor_files/input.txt. You can also create your own files, say random.txt in the predictor_files directory and paste it in there as well.
• Next Open the CLI in repo's home directory.
• cd to Code/ directory. cd Code
• Run the following command to preprocess the raw data. python data_cleaning.py. This will clean the datasets and save them in Cleaned_Datasets/ directory.
• Run the following command if you want to get top 3 recommendations for the resume you input, python model.py.

• You can use the -f <filename> or --filename <filename> options to specify the name of the file in the predictor_files/ directory that you copied your resume contents to.
• The model points to the predictor_files/ directory by default for accessing your resume contents so if you are creating new file then create and copy in that directory only.
• Example command - python -f example1.txt, python --filename example1.txt, note that example1.txt must be in predictor_files directory.
• Also, you may not be from IT industry only, you may be a doctor, architect, etc, the model takes into consideration by default that you are from IT industry but you can specify that you are from NON-IT field by using the -o NON-IT or --option NON-IT flags.
• Example command - to get top 3 recommendations for a NON-IT resume whose contents are pasted in predictor_files/example2.txt we use the command python model.py -o NON-IT -f example2.txt
• Remember that for theses commands to work your current directory must be Job-Recommender/Code/, if its not the same, then cd to it and then use.

• cd Code/
• python data_cleaning.py
• Copy resume contents to predictor_files/input.txt
• Run python model.py