System and method for record identification

A record identification method, comprising:

• at a computing device having a machine learning model operating therein,
• obtaining one or more event occurrence records produced by a surveillance system deployed at one or more monitored locations, the plurality of event occurrence records being indicative of at least one event having previously occurred at the one or more monitored locations and detected by the surveillance system, each of the one or more event occurrence records having a first plurality of attributes associated therewith;
• executing the machine learning model to determine a probability distribution of the first plurality of attributes, the probability distribution of the first plurality of attributes comprising a first vector of first probabilities, each first probability associated with a respective one of the first plurality of attributes;
• receiving one or more witness statements relating to an incident, each of the one or more witness statements having a respective second plurality of attributes descriptive of the incident associated therewith;
• comparing the probability distribution of the first plurality of attributes to a probability distribution of the respective second plurality of attributes, the probability distribution of the second plurality of attributes comprising a second vector of second probabilities, each second probability associated with a respective one of the second plurality of attributes;
• identifying, based on the comparing, at least one of the one or more event occurrence records for which the probability distribution of the first plurality of attributes matches the probability distribution of the respective second plurality of attributes;
• and
• outputting the at least one of the one or more event occurrence records.