def __init__(self,
n_items=30,
reward_func="likelihood",
threshold=0.5,
seed=None,
n_steps=None,
*args,
**kwargs):
super(GPLEnv, self).__init__(n_items=n_items,
reward_func=reward_func,
threshold=threshold,
seed=seed,
*args,
**kwargs)
self.learners = GPLLearnerGroup(self.n_items,
n_steps=n_steps,
seed=seed)
self.test_item_base = self._item_base = MBSItemBase(
self.n_items,
attributes={"difficulty", "difficulty_coefficient"},
seed=seed)
self.learning_item_base = deepcopy(self._item_base)
self.learning_item_base.drop_attribute()
self.test_item_base = self._item_base
self.action_space = ListSpace(self.learning_item_base.item_id_list,
seed=seed)
def __init__(self, seed=None, initial_step=20):
self.random_state = np.random.RandomState(seed)
parameters = load_environment_parameters()
self.knowledge_structure = parameters["knowledge_structure"]
self._item_base = KSSItemBase(parameters["knowledge_structure"],
parameters["learning_order"],
items=parameters["items"])
self.learning_item_base = deepcopy(self._item_base)
self.learning_item_base.drop_attribute()
self.test_item_base = self._item_base
self.scorer = KSSScorer(parameters["configuration"].get(
"binary_scorer", True))
self.action_space = ListSpace(self.learning_item_base.item_id_list,
seed=seed)
self.learners = LearnerGroup(self.knowledge_structure, seed=seed)
self._order_ratio = parameters["configuration"]["order_ratio"]
self._review_times = parameters["configuration"]["review_times"]
self._learning_order = parameters["learning_order"]
self._topo_order = list(nx.topological_sort(self.knowledge_structure))
self._initial_step = parameters["configuration"][
"initial_step"] if initial_step is None else initial_step
self._learner = None
self._initial_score = None
self._exam_reduce = "sum" if parameters["configuration"].get(
"exam_sum", True) else "ave"
def __init__(self,
name,
mode="with_measurement_error",
seed=None,
parameters: dict = None):
"""
Parameters
----------
name:
binary or tree
parameters: dict
* test_item_for_each_skill
mode
"""
super(TMSEnv, self).__init__()
directory = ENV_META.get(name, name)
environment_parameters = load_environment_parameters(directory)
if parameters is not None:
environment_parameters.update(parameters)
self.knowledge_structure = environment_parameters[
"knowledge_structure"]
self.learners = LearnerGroup(
transition_matrix=environment_parameters["transition_matrix"],
state2vector=environment_parameters["state2vector"],
initial_states=environment_parameters["initial_states"],
seed=seed)
self._skill_num = environment_parameters["configuration"]["skill_num"]
self._test_item_for_each_skill = environment_parameters[
"configuration"]["test_item_for_each_skill"]
self.mode_str = mode
self.mode = MODE[mode]
self.learning_item_base = TMSLearningItemBase(
{"d%s" % i: {
"knowledge": i
}
for i in range(self._skill_num)})
self.test_item_base = TMSTestItemBase(self._skill_num,
self._test_item_for_each_skill,
seed=seed)
self.scorer = TMSScorer()
self._learner = None
self._begin_state = None
self.action_space = ListSpace(self.learning_item_base.item_id_list,
seed=seed)
def __init__(self,
n_items=30,
reward_func="likelihood",
threshold=0.5,
seed=None,
*args,
**kwargs):
super(HLREnv, self).__init__(n_items=n_items,
reward_func=reward_func,
threshold=threshold,
seed=seed,
*args,
**kwargs)
self.learners = HLRLearnerGroup(self.n_items, seed=seed)
self.test_item_base = self._item_base = MBSItemBase(self.n_items,
seed=seed)
self.learning_item_base = deepcopy(self._item_base)
self.learning_item_base.drop_attribute()
self.action_space = ListSpace(self.learning_item_base.item_id_list,
seed=seed)