Abstract: A log in which a series of events occurred as a result of the execution of the target program has been recorded is inputted. A plurality of partial logs are created from the inputted log. These partial logs are normalized on the basis of the master log serving as a normalization reference. The feature value representing the degree of feature of the occurrence and nonoccurrence of an event is calculated for each of the normalized logs on the basis of the normalized logs for the remaining partial logs. In a combination of a specific partial log and another partial log, the similarity between these partial logs is calculated by performing an operation on the basis of the feature values. For example, a combination of the specific partial log and the partial log with the highest similarity is displayed.

Abstract: One of log analysis methods of the present invention includes the step of executing a program a plurality of times; executing a program a plurality of times; the step of generating a plurality of logs, each log being recorded a plurality of events occurring upon the execution of the program according to an occurrence order of each of the events in each of the logs; the step of performing a first calculation to calculate an event occurrence probability, for the occurrence order of each event, based on at least one from the program description concerning each event recorded in the logs and the data to be used upon the execution of the program description; and the step of outputting information concerning an event which corresponds to a characteristic included in the logs, based on the event occurrence probability.

Abstract: One of log analysis methods of the present invention comprises the step of executing a program a plurality of times; executing a program a plurality of times; the step of generating a plurality of logs, each log being recorded a plurality of events occurring upon the execution of the program according to an occurrence order of each of the events in each of the logs; the step of performing a first calculation to calculate an event occurrence probability, for the occurrence order of each event, based on at least one from the program description concerning each event recorded in the logs and the data to be used upon the execution of the program description; and the step of outputting information concerning an event which corresponds to a characteristic included in the logs, based on the event occurrence probability.

Abstract: A log in which a series of events occurred as a result of the execution of the target program has been recorded is inputted. A plurality of partial logs are created from the inputted log. These partial logs are normalized on the basis of the master log serving as a normalization reference. The feature value representing the degree of feature of the occurrence and nonoccurrence of an event is calculated for each of the normalized logs on the basis of the normalized logs for the remaining partial logs. In a combination of a specific partial log and another partial log, the similarity between these partial logs is calculated by performing an operation on the basis of the feature values. For example, a combination of the specific partial log and the partial log with the highest similarity is displayed.