Projects
Projects
In the fast-paced world of business, numerous phone calls and meetings occur where critical decisions are made by various participants. Traditionally, someone would manually take notes during these discussions and distribute them to attendees to confirm the outcomes. By leveraging voice-to-text technologies and implementing AI capable of understanding the conversation’s context, it is possible to automate the meeting summary process efficiently.
Voice-Text Conversion
Multiple Speaker Detection
Multiple Language Detection and Translation
Extractive Text Summarisation Ensembling using NLP PageRank and Sentence Embeddings.
Meeting/ Phone Analytics.
Minutes of the meeting are sent using EMAIL to the participants.
Web APP - HTML, CSS, JS, Flask, Python
Voice -> Text/Multiple Speaker Detection - IBM Watson Speech - text -API, REST
Multiple Language Detection and Translation - Google Translate
Extractive Text Summarisation Ensembling using NLP PageRank and Sentence Embeddings - Numpy, NLTK
Meeting Analytics - Pandas, d3, Gensim, word cloud, spacy, Topic Modelling using LDA.
Thousands of ICOs(initial coin offerings are a form of cryptocurrency that businesses use to raise capital) are launched every year with white papers outlining their vision and strategy. A massive amount of these projects are scams designed to steal money from investors. Distinguishing between credible projects from scams or poorly conceived ideas is critical for investors and analysts. This project leverages state-of-the-art natural language processing and deep learning-based AI models to efficiently automate the classification of ICO white papers. The classifier identifies patterns to determine a project's legitimacy or focus area by analyzing context, technical terminology, sentiment, and readability. This automated approach streamlines the evaluation process, ensuring stakeholders can make informed decisions swiftly and accurately.
Web Scraping and White Paper Pre-Processing: Scrape Icobench.com for new cryptocurrencies and get their whitepaper.
Contextual Classification: Distinguishes between legitimate, fraudulent, and technology-driven ICOs.
Sentiment Analysis: Detects overly optimistic or suspicious language patterns.
Keyword Extraction: Identifies blockchain-specific terminology and critical project details.
Automated Evaluation: Reduces the need for manual analysis, speeding up decision-making.
Web Scraping: Beautiful Soup, and Python for web scraping.
NLP Models: Deep-learning-based models for classifications and contextual understanding.
Data Processing: Numpy, Pandas, NLTK and spaCy for feature extraction and text preprocessing.
Classifier: Assigns a risk percentage for the whitepapers for better analysis and investment decisions.