🔎 Technical Overview
The Token Scoring System evaluates and scores crypto tokens on the Solana blockchain based on three key dimensions:
Twitter Sentiment and Engagement: Analyzes social media activity for sentiment, popularity, and trends.
Smart Contract Analysis: Examines token contracts for safety, functionality, and risk factors.
Historical Data Analysis: Evaluates historical performance metrics, adoption rates, and market behavior.
The system integrates insights from all three dimensions into a unified scoring model to provide a comprehensive evaluation of each token.
System Architecture
1. Data Collection
Twitter: Tweets are collected using the Twitter API, filtered by token-related keywords and hashtags.
Smart Contracts: Contract data is fetched using Solana’s RPC endpoints and parsed for features.
Historical Data: Market and blockchain data is ingested via APIs and stored in a database.
2. Feature Engineering
Twitter:
Sentiment Analysis: Uses a fine-tuned RoBERTa model to classify sentiment.
Topic Modeling: Uses BERTopic to group tweets by themes.
Engagement Scoring: Calculates weighted scores based on likes, retweets, and replies.
Smart Contracts:
Contract Safety: Anomaly detection using Isolation Forests.
Functionality Scores: Flags contracts with unusual or high-risk patterns.
Network Analysis: Graph-based analysis using GraphSAGE.
Historical Data:
Time-Series Trends: Uses LSTMs to model price and volume trends.
Behavioral Clustering: Groups tokens with similar performance patterns using k-Means.
Predictive Models: Regression models (e.g., XGBoost) predict future performance.
3. Model Integration
Multi-Modal Fusion:
Combines Twitter, contract, and historical insights using a weighted ensemble model.
Assigns weights to each dimension (e.g., 40% Twitter, 30% contract, 30% historical data).
4. Scoring
Composite Score: Produces a score between 0-100, with higher scores indicating better overall token performance.
Last updated
Was this helpful?