Update TriRank documentation

[lthoang]

Description

Related Issues

https://github.com/PreferredAI/cornac/issues/618

Checklist:

• [ ] I have added tests.
• [ ] I have updated the documentation accordingly.
• [ ] I have updated README.md (if you are adding a new model).
• [ ] I have updated examples/README.md (if you are adding a new example).
• [ ] I have updated datasets/README.md (if you are adding a new dataset).

[tqtg]

@lthoang there is a failed unit test related to NextItemRecommender. See the log below:

_______________________ TestNextItemRecommender.test_fit _______________________

self = <models.test_recommender.TestNextItemRecommender testMethod=test_fit>

    def test_fit(self):
        dataset = SequentialDataset.from_usit(self.data)
        model = NextItemRecommender("")
        model.fit(dataset)
        model = SPop()
        model.fit(dataset)
        model.score(0, [])
        result = model.rank(0, history_items=[])
>       self.assertTrue((result[0] == [3, 2, 4, 1, 0, 5, 8, 7, 6]).all())
E       AssertionError: False is not true

tests/cornac/models/test_recommender.py:88: AssertionError

There might be some randomness in this test where it passes most of the time but fails rarely?

[lthoang]

@lthoang there is a failed unit test related to NextItemRecommender. See the log below:

_______________________ TestNextItemRecommender.test_fit _______________________

self = <models.test_recommender.TestNextItemRecommender testMethod=test_fit>

    def test_fit(self):
        dataset = SequentialDataset.from_usit(self.data)
        model = NextItemRecommender("")
        model.fit(dataset)
        model = SPop()
        model.fit(dataset)
        model.score(0, [])
        result = model.rank(0, history_items=[])
>       self.assertTrue((result[0] == [3, 2, 4, 1, 0, 5, 8, 7, 6]).all())
E       AssertionError: False is not true

tests/cornac/models/test_recommender.py:88: AssertionError

There might be some randomness in this test where it passes most of the time but fails rarely?

Investigate this failing case, I find that the ranking results of SPop as follows:

(array([3, 2, 4, 0, 1, 5, 8, 7, 6]), 
array([0.6923077 , 0.6923077 , 0.9230769 , 1., 0.7692308 , 0.3846154 , 0.07692308, 0.07692308, 0.07692308], dtype=float32))

We can see that there are some items having identical scores. Thus we need to rewrite to ensure the test case can pass in those cases. This will be fixed in another PR.