Some basic concepts of Data Mining

Modeling Method:

1: Data Summary:

Eg: PageRank

Data is used to reflect the importance of a Web page, that is, the probability that a random user is on this page.

2: Clustering

3. Feature Extraction

1: frequent itemset:

Eg: most users buy A and B at the same time. When a user buys a, B is recommended to him.

2: similarity item similar item: Collaborative Filtering

Search for similar users or products for recommendation

The theory of bonefrani: positive correlation between pancreatic production, that is, when a specific feature is searched for in certain data, even if the data is completely random, the feature will appear and increase as the data grows.

TF. IDF: TF * IDF

TF: Word Term Frequency

TF (ij) = f (ij)/maxk F (kJ)

That is, in document J, the frequency of word item I is the frequency of word item I divided by the frequency of word item k with the highest frequency (normalization without considering deprecated words)

IDF: Frequency of inverse document

Assume that the number of documents is N. If word item I appears in N (I) documents, the IDF of word item I is:

IDF (I) = Log (2) n/n (I)

Not complete...