💶Kafka-SparkStreamNLP 是一个基于docker容器化管理的实时金融文本分析平台，通过新闻api，采用 Kafka 进行数据流管理，使用 Spark Streaming 结合微调预训练模型finetuning进行NLP处理，并通过输出流将结果存储在clickhouse以便后续使用可视化平台进行olap分析⭐️⭐️⭐️⭐️⭐️

Academic Sequence Labelling Between DistillBERT & Encoder-only Transformer

Extrinsic and Intrinsic Plagiarism detection

Work focus on Transformer model to Start classification (1-5) about reviews of YELP.

Advanced RoBerta and DistillBert Based Abstract Based Sentiment Analyzer. Ensemble Architecture

Sentiment Analysis

Distilbert model for sentence segmentation.

End-to-end backend system for extracting structured data from documents using OCR, transformer-based NER, and asynchronous processing with FastAPI.

AI-Enhanced Exploratory Analysis and Preprocessing of Simulated Patient Vital Signs Data

Streamlit-based Business Analytics System for call sentiment analysis, audio transcription, and live prediction using ML and NLP.

A mobile app that uses Machine Learning to detect and block spam or scam texts.

Movie review sentiment analysis by using DistillBERT

PR/Issue Classifier - ML model to automatically label GitHub PRs and issues, predict sentiment, and set priority

A Hybrid Quantum-Classical Machine Learning framework using DistilBERT and Variational Quantum Circuits (VQC) for advanced binary fake news detection on the LIAR dataset.

Classification of Text from Youtube Comments using BistillBERT alanguage models from Hugginface Transformers

AI-powered medical symptom checker using DistilBERT (Small Language Model) to predict possible diseases from user-reported symptoms with confidence visualization.

Hybrid Extractive-Abstractive Financial Text Summarization using Semantic Clustering, DistilBERT Sentence Scoring, and Longformer Encoder-Decoder (LED) for long-document summarization of SEC Form 10-K reports.

NLP pipeline for sentiment analysis and BERTopic topic modelling on 212,000+ Google Maps restaurant reviews.

A machine-learning project that analyzes the sentiment of tweets using deep learning and NLP techniques. The model classifies tweets into positive or negative sentiment, using preprocessing, tokenization, and training on a labeled dataset. Includes data cleaning, visualization, model training, evaluation (accuracy, precision, recall, F1-score),

Small NLP projects with Deep Learning techniques