A graph algorithm for semantic map models

Functions

1. To construct a semantic map according to the instance-function matrix.

2. To visualize the map using nodes and weights.

3. To evaluate the map in terms of internal statistics, the requirements of continuity hypothesis, and the differences between the ground truth.

Results

1. Our method can get a competitive result according to the evaluation.
2. Our method is far more efficient than human constructions.

Examples

1. A dialect investigation to “敆”
2. Repetitive adverbs
3. Repetitive adverbs with more functions

Several Statistics (Update: 2025.01.13)
Precision: 0.0015100037750094375         Recall: 1.0     F1: 0.003015454202488204
Summed Weight: 102
Network typology of degree mean: 2.111111111111111       std: 1.9116278371205837

ACC_GT: 0.9382716049382716

Contacts

Feel free to contact us if you have any ideas about the examples, applications and so on. You can just pose an issue or drop me email by liuzhu22@mails.tsinghua.edu.cn

Notes

This is still an ongoing project. More features are coming soon!