Risk, confidence and coverage answer three different questions

Observed risk is not a fraud probability. Confidence and coverage provide essential context for interpreting a finding.

Risk describes the signals observed

In Fake Review Detector, risk describes the warning signals established from the material examined. It helps surface issues that may deserve attention.

It should not be interpreted as proof of fraud or a probability that a business is fraudulent. An assessment still needs its underlying observations and the limitations of those observations.

Confidence describes the support for a finding

Confidence asks how strongly the available evidence supports an observation or interpretation. Clear, traceable material can support a finding more directly than incomplete or ambiguous material.

Confidence is not the same as safety. A well-supported observation can be narrow, and a weakly supported concern may need further investigation before it becomes useful.

Coverage describes what was examined

Coverage asks how much of the relevant material was available and checked. A report may examine a particular review sample or a set of website pages while leaving other questions unanswered.

High confidence in one observation cannot fill a coverage gap elsewhere. This is why unknowns, inaccessible pages and missing information should remain part of the report.

Read the three together

An assessment becomes more useful when these ideas stay distinct. Instead of reducing everything to one reassuring number, ask what each finding can actually support.

• What signal was observed?

• Which evidence supports it?

• How clear is that evidence?

• What material was checked?

• Which questions remain open?

Choose the next question, not a final label

Use the report to decide where more information would matter. You may need another source, a clearer description of a claim or a broader sample of observations.

The aim is to make uncertainty understandable so that a decision has more context. Automated analysis supports that process; it does not replace the judgement required to interpret the evidence.