-
About Us+
-
Products+
-
Solution+
-
Cases+
-
AI Lab+
-
Investor+
-
Contact+
-
Join Us +
-
-
Trial+
About Us+
Products+
Solution+
Cases+
AI Lab+
Investor+
Contact+
Join Us +
Trial+
This paper mainly introduces an efficient unsupervised algorithm isolated forest in the field of outlier detection: the algorithm constructs multiple isolated trees, divides each sample into a separate category, and finds outliers by measuring the difficulty of distinguishing each sample point during classification. In the past, most outlier detection methods used supervised models, which need to use historical data to classify transactions. Therefore, these models can only identify the existing fraud methods in historical data, but new fraud methods are difficult to identify. The unsupervised detection principle of isolated forest can overcome this problem, and the algorithm also has the characteristics of high precision and linear time complexity, which is very suitable for the anti fraud field of financial industry.
Download