There are various measures of performance – accuracy, recall, precision etc. The performance is determined against human labelled data as true positive. There’s also ‘folding’ to validate models for stability.
Articles in this section
- How do I use the labels and metrics that PrediCX generates?
- How do I import data automatically?
- Do I need to be a data scientist to use PrediCX?
- Can I change the labels later on, or are they set in stone from day one?
- Can I have multiple labels per record?
- Can I use PrediCX on any text dataset, and will it work even with complex or technical text?
- How does PrediCX measure its' performance?
- How do I generate labels to start with?