Research Article
Gender Demographic Prediction on Twitter: A Comprehensive Review of Machine Learning and Data Analytics Approaches
@INPROCEEDINGS{10.4108/eai.19-12-2025.2366281, author={Tarun Kumar Sharma and Vishal Shrivastava}, title={Gender Demographic Prediction on Twitter: A Comprehensive Review of Machine Learning and Data Analytics Approaches}, proceedings={Proceedings of the 1st International Conference on Computational Techniques in Data Science, IICTDS 2025, 19-20 December 2025, Chandigarh, India}, publisher={EAI}, proceedings_a={IICTDS}, year={2026}, month={11}, keywords={Gender prediction Twitter analytics machine learning deep learning demographic profiling natural language processing feature engineering social media mining SVM Random Forest XGBoost BERT transfer learning data imbalance}, doi={10.4108/eai.19-12-2025.2366281} }- Tarun Kumar Sharma
Vishal Shrivastava
Year: 2026
Gender Demographic Prediction on Twitter: A Comprehensive Review of Machine Learning and Data Analytics Approaches
IICTDS
EAI
DOI: 10.4108/eai.19-12-2025.2366281
Abstract
In this paper, the author will discuss the design and analysis of cutting-edge machine learning (ML) and data analytics solutions that will help increase the accuracy and reliability of gender prediction using Twitter data. The paper critically analyzes various state-of-the-art models such as Support Vector Machines (SVM), Random Forest, Gradient Boosting algorithms (XGBoost, LightGBM), and deep learning models such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs) and Transformer-based models (BERT, RoBERTa). Particular attention is given to such preprocessing methods as text normalization, tokenization, feature engineering (n-grams, POS tagging, embeddings), and incorporation of multimodal information - the textual information, profile metadata, user activity patterns. Moreover, the review also emphasizes the importance of dataset balancing with such methods as SMOTE and the influence of the linguistic, cultural, and regional diversity on the generalization of the model. The relative analysis of the benchmark data sets, including the PAN Author Profiling, indicates the main trends of performance and efficiency of the ensemble learning and transfer learning methods.