Baseline + experiments

I created a pipeline for the system:

A baseline + 23 variants of the baseline (experiments with different preprocessing steps, different clustering algorithms, different language models.)

The zip.document consists of

• 24 .py files with 24 implementations of baselines
• 24 output files in the right format
• 24+24 evaluation reports
• a table with results of experiments

Baselines.zip

The performance doesn't differ very much for all the variants, that's why I want to make a simple test with a small toy example to check the implementation + a simple test fot evaluation to check it.

As we do induction, our labels for clusters do differ from the labels for clusters in the gold standard, that's why it's important to understand how it influences the performance

closed