8 Anomalies Limits

BigML anomalies have a few limitations regarding the type of input data they can support. BigML also impose some limits on the configurable parameters to create an anomaly. You can find all limits listed below:

• Categorical fields: a maximum number of 1,000 distinct classes per field is allowed.

• Text fields: anomalies do not support text as input fields.

• Item fields: anomalies do not support items as input fields.

• Top Anomalies: a maximum of top 1,024 anomalies is allowed. You can score all your dataset instances by performing a batch anomaly score.

• Forest size: a maximum of 256 trees for the Isolation Forest is allowed.