Dr. Suruchi Sabherwal

T-Bot: AI-based Bot Detection Framework for Trend-centric Online Social Networks
This research presents T-Bot, an AI-driven framework designed to detect malicious bots in trend-centric online social networks such as Twitter/X. The framework integrates machine learning techniques with temporal and network-based features to identify abnormal behavior patterns during trending events. Unlike traditional bot detection approaches, T-Bot focuses on trend dynamics, enabling more accurate detection of coordinated and automated activities. Extensive experiments demonstrate improved precision and recall compared to existing baseline models. The study contributes to strengthening trust, security, and authenticity in online social platforms by mitigating automated manipulation.

---

C-ANN: A Deep Learning Model for Detecting Black-Marketed Colluders in Online Social Networks
This paper proposes C-ANN, a deep learning-based model aimed at detecting black-marketed collusive users in online social networks. The model leverages artificial neural networks to capture hidden behavioral patterns of colluders who artificially boost popularity metrics such as likes, retweets, and followers. By combining user activity features with network interaction data, the approach achieves high detection accuracy. The work addresses a critical challenge in social media integrity and provides a scalable solution for identifying coordinated inauthentic behavior in large-scale networks.

---

Black Marketed Collusive Users Primary Dataset from Twitter/X Online Social Media
This publication introduces a publicly available primary dataset containing black-marketed collusive user activity from Twitter/X. The dataset is curated to support research in social network analysis, bot detection, and malicious user behavior modeling. It includes user metadata, interaction patterns, and temporal features useful for machine learning and graph-based analysis. By providing a benchmark dataset, this work enables reproducibility, comparative evaluation, and advancement of research in detecting social media manipulation and coordinated campaigns.

---

A Machine Learning-based Malicious Bot Detection Framework for Trend-centric Twitter Stream
This study proposes a machine learning framework for identifying malicious bots operating within trend-centric Twitter streams. The framework extracts content-based, temporal, and user-level features to classify automated accounts during trending discussions. The proposed approach significantly outperforms traditional detection methods and was recognized with a Best Paper Award. The work contributes to real-time social media monitoring and cyber security by addressing emerging threats in online communication ecosystems.

---

Research in data science, machine learning, and network science with focus on large-scale online social networks.

---

Specialized in bot detection, collusive user identification, and malicious behavior analysis on social media platforms.

---

Developed AI and deep learning–based frameworks for trend-centric social network analysis.

---

Published in reputed SCI/Scopus-indexed journals and IEEE conferences.

---

Applied graph analytics and explainable AI for influence and interaction modeling.

---

Guided student research projects addressing real-world data science and cyber security problems.

---

Research Interests

---

Data Science and Applications

---

Machine Learning and Deep Learning

---

Network Science and Social Network Analysis

---

Online Social Media Analytics

---

Bot Detection and Collusive User Identification

---

Big Data Analytics

---

Large-Scale Graph Algorithms

---

Explainable AI (XAI)

---

Cyber Security and Social Media Crimes

---

Data Visualization and Interpretation

---

Best Paper Award – International Conference on Networks and Cryptology (NetCrypt 2020)
Received Best Paper Award for research on a machine learning–based malicious bot detection framework for trend-centric Twitter streams.

---