Web3AI

By tracking the latest updates of over 20,000 Web3 projects and 50,000 KOLs, and combining AI big data analysis, we uncover the social trends and popularity of each project to guide users in their investment decisions.

Utilizing AI sentiment algorithms, we keep abreast of the latest market dynamics in real time, continuously monitoring the latest tweets from all KOLs, recording and comparing daily sentiment scores and price trends, and analyzing changes in mentions relative to price fluctuations. Users can directly view the overall daily Twitter sentiment, which serves as a critical indicator for assessing the current market mood.

Today's KOL Overall Sentiment Index = Total Project Sentiment Scores / Number of Projects

AI Sentiment Model Training Process

• Data Collection

  • Using the Twitter API: Collect or scrape over 50,000 tweets related to Web3 industry key opinion leaders (KOLs), using keywords and thematic tags such as #Web3 and #Blockchain.

  • Real-time Monitoring: Establish a real-time data stream to analyze current sentiment shifts.

• Data Preprocessing

  • Data Cleaning: Remove noise such as URLs, special characters, and user mentions.

  • Language Standardization: Address spelling errors and slang to standardize the text.

• Model Selection and Training

  • Pre-trained Models: Utilize pre-trained models like RoBERTa or BERTweet, which can be further fine-tuned on a Web3-specific dataset.

  • Sentiment Classification: Employ existing sentiment classification models (e.g., positive, neutral, negative) and enhance accuracy through fine-tuning on Web3-specific data.

• Fine-tuning and Evaluation

  • Data Labeling: Manually label a subset of Web3 tweets for the training set.

  • Model Fine-tuning: Perform fine-tuning on the labeled dataset.

  • Performance Evaluation: Assess the model using metrics such as accuracy, recall, and F1 score.

• Deployment and Application

  • API Deployment: Deploy the model as an API for application integration.

  • Real-time Analysis: Integrate into dashboards to display real-time sentiment trends.

• Visualization and Reporting

  • Sentiment Trend Graphs: Display changes in sentiment over time.

  • Word Clouds: Show the most frequently used positive and negative words.

• Feedback and Optimization

  • User Feedback: Gather user feedback to refine the model.

  • Continuous Learning: Regularly update the model, incorporating the latest tweet data for retraining.