decodIRT consolidated version

Consolidating current version of decodIRT after fixing some bugs and adding some exception handling. Preparing for a major tool update.

decodIRT as package and publish benchmark in OpenML

• The decodIRT tool has been modified to work as a package as well. And it allows your scripts to be executed as a module in python code.

• New features have also been added. The new module decodIRT_create_benchmark, allows you to create a benchmark using the IRT data generated by the decodIRT and publish the benchmark in OpenML as a study. Only using a few lines of code.

IRT score implemented (BRACIS20)

• Finished the functions for calculating the IRT score.
• Added functions to plot the comparative graph of the scores.

Initial aptitude test implementations

• Modified as the classifier proficiency is calculated. Now it's gradual (from easiest to most difficult)
• Added new outputs for decodIRT_analysis (theta list, score based on probability of hit)

All scripts are functional

All scripts are ready for use