machines.zip contains the set of state machines in GraphViz dot format. Some of these have had to be converted from other formats.
ClassificationResults.zip contains the raw data returned by the Mint inference tool for each subject system. For each machine this produces two files:
One file with the suffix _inference, which contains the raw data. The heading for this file is:
Model, Data_inference_algorithm, random_seed, k-tails, infer_data, merging_algorithm, alpha, belief, disbelief, uncertainty, predicted, actual, num_states, num_transitions
These are defined as follows:
* Model - name of the model from which traces have been sampled.
* Data_inference_algorithm - remains constant at J48 and is ignored
* random_seed - the random seed used to infer the model.
* k-tails - remains constant at 0.
* infer_data - remains constant (false) - there are no data variables in the traces.
* merging_algorithm - remains constant at 'redblue'.
* alpha - the uncertainty threshold set - either 250, 500, 1000, or 2000.
* belief - the belief for the prediction for the given row.
* disbelief - the disbelief for the prediction for the given row.
* uncertainty - the uncertainty for the prediction for the given row.
* predicted - does the inferred model predict an accept or reject
* actual - is the correct response an accept or reject
* num_states - number of states in the inferred model
* num transitions - number of transitions in the inferred model
The file with the suffix "inference stats" contains some stats for the target model. The heading (which is quite self-explanatory) is:
num_states, num_transitions, alphabet_size
the file SOSMScripts.rmd is an R Markdown language, which contains the various analyses that were used to derive the statistics and charts for the paper from the classification results.
mint.jar is included for completeness - this is the version of MINT that was used.