Resume-Selector

Overview

Resume-Selector is a machine learning project designed to analyze and classify resumes based on industry sectors. Using a dataset from Kaggle, this project employs various preprocessing techniques and machine learning algorithms to streamline the hiring process.

Features

• Data Cleaning: Remove special characters, stopwords, and irrelevant information from resumes.
• Feature Extraction: Utilize TF-IDF to convert textual data into numerical vectors suitable for machine learning.
• Classification: Implement a machine learning model to categorize resumes by industry.

Dataset Description

The dataset for this project was sourced from Kaggle and is in Excel format, consisting of three columns:

• ID: Sequence number of the resume
• Category: Industry sector to which the resume belongs
• Resume: The complete CV of the candidate

Preprocessing Steps

The preprocessing of resumes involves several key steps to ensure data quality:

• Data Cleaning:

  • Remove special characters, single-letter words, and unique letters/numbers from resumes.
  • Tokenize the text using the NLTK tokenizer.

• Masking:

  • Mask strings using \w.
  • Mask escape characters like \n.
  • Mask all numbers.
  • Remove email addresses.

• Stopword Removal:

  • Common stopwords (e.g., "and", "the", "was") are removed as they limit prediction accuracy.
  • Input words are tokenized into individual tokens and checked against the NLTK stopwords list:

    from nltk.corpus import stopwords
    stop_words = set(stopwords.words('english'))

  • Words found in the stopwords list are removed from the main array until no stopwords remain.

• Feature Extraction:

  • The TF-IDF (Term Frequency-Inverse Document Frequency) method is used to extract features from the preprocessed dataset.
  • Cleaned data is transformed into numerical vectors for classification.
  • Utilized TfidfVectorizer from the Scikit-learn library with the following configurations:
    • Sub-linear DF: Set to True to apply a logarithmic scale for frequency.
    • Min DF: Defines the minimum number of documents a word must appear in to be included.
    • Norm: Set to L2 to ensure all feature vectors have the same Euclidean norm.
    • N-gram Range: Set to (1, 2) to include both unigrams and bigrams.

By following these steps, the project effectively prepares the resumes for accurate analysis and classification.

Dataset

The dataset used in this project is sourced from Kaggle and is in Excel format. It includes the following columns:

• ID: Sequence number of the resume
• Category: Industry sector associated with the resume
• Resume: Full text of the candidate's CV

Technologies Used

• Python
• NLTK for natural language processing
• Scikit-learn for machine learning
• Pandas for data manipulation
• NumPy for numerical operations

Installation

• Clone the repository:

  git clone https://github.com/AyushAggarwal1/Resume-Selector.git
  cd Resume-selector
  

Report (Results)

Research Paper