Visit @ github

Description

WordNet is a semantic lexicon project that implements a graph-based system for analyzing relationships between English nouns. The project involves constructing a directed acyclic graph (DAG) from WordNet's semantic database, where vertices represent synsets (groups of synonymous words) and edges represent "is-a" relationships between concepts. This implementation focuses on efficiently processing semantic relationships between nouns and identifying semantic outliers within groups of words.

Features

• Graph Construction and Data Management
  • Constructs a WordNet directed graph from input files with linear time and space complexity
  • Implements efficient bidirectional mapping between nouns and synset IDs
  • Provides optimized lookups for both noun-to-synset and synset-to-noun relationships
• Semantic Analysis Capabilities
  • Calculates shortest common ancestors (SCA) between synset pairs
  • Determines semantic distance between noun pairs
  • Processes multiple word meanings through synset relationships
  • Handles complex cases where nouns appear in multiple synsets
• Query Processing
  • Computes semantic relatedness between noun pairs
  • Detects semantic outcasts in noun groups
  • Performs efficient path analysis in the semantic network
  • Maintains performance optimization for large-scale queries

Outcomes

• Performance Metrics
  • Graph Construction: Linear time and space complexity relative to input size
  • Noun-Synset Mapping: Sub-linear time complexity (better than O(N))
  • Distance Calculations: O(N + M) complexity, where N is number of synsets and M is number of is-a relationships
• Functional Deliverables
  • Accurate semantic relationship measurements between noun pairs
  • Precise identification of semantic outcasts in noun groups
  • Efficient shortest common ancestor detection
  • Optimized path finding in the semantic network
• System Outputs
  • Semantic relatedness metrics for noun pairs
  • Distance measurements between concepts
  • Shortest common ancestor identification
  • Outcast detection results for noun groups

Stack