Contents
1 Introduction and Theoretical Background
1.1 Multiple Context Free Grammars
1.2 MCFG String Parsing
1.3 Prefix Parsing as Intersection of (M)CFG and Finite State Automaton
1.4 Probabilistic Parsing and MCFG
1.5 Probabilistic Intersection of a (M)CFG and Finite State Automaton
1.5.1 Inside Probability
1.5.2 Renormalization by Inside Probability
1.6 Entropy
2 Quick User Guide to mcfgcky
2.1 Debugging Mode
2.2 Parsing Mode
2.3 Statistical Mode
3 Design
3.1 Chart Items
3.2 Inference Rules
3.3 Agenda and Exhauastion Mechanism
3.4 Backpointing
3.5 Output
3.6 Training
3.7 Testing
3.8 Inside
3.9 Surprisal
3.10 Entropy
4 Implementation
4.1 Utilities, Basic Data Structures, and Input Modules
4.1.1 Module
Utilities
4.1.2 Module
MyQueue
4.1.3 Module
Seq
4.1.4 Module
Mcfgread
4.1.5 Module
Fileread
4.1.6 Module
Corpread
4.1.7 Module
Pcfgead
4.2 Parsing Modules
4.2.1 Module
Item
4.2.2 Module
Grammar
4.2.3 Module
Point
4.2.3.1 Experimental Kilbury Parsing
4.2.4 Module
Concat
4.2.4.1 Integrity Checks
4.2.4.2 Sanity Concatenation Checks
4.2.4.3 Concatenation Functions
4.2.5 Module
Rules
4.2.5.1 Chart Population Functions
4.2.5.2 Inference Rules
4.2.5.3 Ground Axia/Initial Chart Population
4.2.5.4 Kilbury Style Inference Rules
4.2.6 Module
Mcfgcky
4.2.6.1 Experimental Kilbury Stuff
4.3 Statistical Parsing
4.3.1 Module
Pcfg
4.3.2 Module
Train
4.3.3 Module
Inside
4.3.4 Module
Test
4.3.5 Module
Entropy
4.3.5.1 Stolcke matrix cut trick
4.4 Printing and Output
4.4.1 Module
Decompile
4.4.2 Module
Output
4.4.3 Module
Print