DIG4EL General Agent testing result on canonical word order on 116 languages

These files enclose the detailed testing results of a DIG4EL General Agent (GA) test. The GA has been tasked to infer the value of 8 unknown parameters pertaining to word order in 116 languages, given the value of 7 other parameters pertaining to word order that can be automatically observed by the software. A Leave-One-Out method has been used, whereby each test on a language uses a language model that does not include it. Each language is tested over 11 independent epochs. The archive contains all the raw outputs of the software: Truth tables for each epoch of each language. For each parameter, a 0 indicates a failure to predict the correct value, and a 1 indicates a success. This data can be reproduced using the general_agents_quantitative_comparative_analysis.py script in the source code.