AcademiQ: A Computer Science Research Search Engine

AcademiQ is a specialized search engine that helps researchers, students, and institutions find computer science-related papers, authors, and research institutions using Learning to Rank techniques.

---

🚀 Project Overview

The goal is to build a search engine that:

• Identifies scholars and research institutions for computer science topics.
• Optimizes search relevancy using LambdaMART Learning to Rank (L2R) models.

---

🛠️ Core Features

1. Data Integration:
   • Collects data from Google Scholar API, Scopus API, arXiv, and DBLP.
2. Machine Learning:
   • Implements LambdaMART (L2R) for search ranking.
   • Supports LightGBM for fast and efficient training.
3. User Interface:
   • Built with Flask (backend) and Bootstrap (frontend) for user-friendly interaction.

---

🧩 Project Structure

AcademiQ/
├── dataset/            # Data ingestion and preprocessing
├── files/              # LambdaMART and LightGBM code
├── static/             # Project logo and CSS/Bootstrap of template
├── templates/          # HTML template of user interface
├── main.py             # Backend server using Flask
└── README.md           # Project documentation

⚙️ Technical Details

Data Collection

• APIs: Google Scholar, Scopus, arXiv, DBLP.
• Data Includes:
  • Papers (title, abstract, citations).
  • Authors (name, affiliation, citation count).
  • Institutions (organization and rankings).

Learning to Rank (L2R)

• Algorithm: LambdaMART (RankNet + NDCG).
• Framework: LightGBM.
• Training:
  • Features extracted include:
    • TF-IDF, BM25, PageRank, HITS (Hub & Authority scores), Citation count, etc.

Evaluation

• Metric: NDCG@K (Normalized Discounted Cumulative Gain).

---

🖥️ Deployment

• Backend: Deployed on AWS EC2.
• Frontend: Flask handles UI rendering.