A python library of IRT algorithm designed to cope with sparse data structure.

• built and test under py3.6. Python 2 compatibility is tested but not guaranteed.

When install from github source code

pipenv --three
make

from pyirt import irt

src_fp = open(file_path,'r')

# alternatively, pass in list of tuples in the format of [(user_id, item_id, ans_boolean)]
# ans_boolean is 0/1.


# (1)Run by default
item_param, user_param = irt(src_fp)

# (2)Supply bounds
item_param, user-param = irt(src_fp, theta_bnds = [-4,4], alpha_bnds=[0.1,3], beta_bnds = [-3,3])

# (3)Supply guess parameter
guessParamDict = {1:{'c':0.0}, 2:{'c':0.25}}
item_param, user_param = irt(src_fp, in_guess_param = guessParamDict)

When dealing with big data, the memory limit of the single machine is usually the bottle neck.

pyirt ships with a pymongo integration that can handle millions of record (we tried 1 billion).

The mongo db connection config is in "settings.ini", whose format is the same as "settings.ini.example"

For usage, see

python -m unittest tests.test_dao.TestDataSrc.test_from_mongo

See wiki

The algorithm is described in details by Bradey Hanson(2000), see in the literature section. I am grateful to Mr.Hanson's work.

Chaoqun Fu's comment leads to the (much better) API design.

Dawei Chen and Lei Wang contributed to the code.