Granted Patent B2

US 12,339,733 · App. 18/180,485 · Granted Jun 24, 2025

Inventors: Ken Tonari (Kanagawa, JP); Ryo Suzuki (Kanagawa, JP)

Assignee: NEC Platforms, Ltd.

G06F11/079G06F11/0793

Abstract

Claims (36)

1. An information providing system comprising:

a memory storing instructions; and

a processor configured to execute the instructions to:

acquire log data generated by an information equipment system;

calculate a first term frequency-inverse document frequency (TF-IDF) vector to extract a feature of the log data;

calculate a second TF-IDF vector as a feature of information included in history information related to the information equipment system;

acquire information from the history information based on a degree of similarity between the first TF-IDF vector and the second TF-IDF vector;

output the acquired information;

generate a model for calculating likelihood of classifying a log for each type of failure and for each time interval acquired by dividing a time until a failure occurs; and

detect a sign of failure occurrence of the information equipment system by using the model and the extracted feature of the log data.

2. The information providing system according to claim 1 , wherein the processor is configured to execute the instructions to extract the feature of the log data from a predetermined time before occurrence of a failure in the information equipment system to a time of the failure.

3. The information providing system according to claim 1 , wherein the processor is configured to execute the instructions to estimate a suspected failure component of the information equipment system based on the feature of the log data.

4. The information providing system according to claim 1 , wherein the processor is configured to execute the instructions to estimate a failure occurrence event of the information equipment system based on the feature of the log data.

5. An information providing system comprising:

a memory configured to store instructions; and

a processor configured to execute the instructions to:

acquire log data generated by an information equipment system;

extract a feature of the log data;

acquire, from history information related to the information equipment system, information based on the feature;

output the acquired information;

determine, among correct classes included in learning data in which past failure logs and the correct classes are linked, whether there is a correct class having a number less than a predetermined number;

change a collection time period of the past failure log linked with the correct class having the number less the predetermined number;

copy the past failure log that the collection time period of which has been changed;

add the copied failure log in the learning data; and

cause a model to learn classification of the extracted feature of the log data by using the learning data.

6. The information providing system according to claim 1 , wherein the processor is configured to execute the instructions to:

add an inverse vector of a feature vector of a normal system log to a feature vector of a failure log included in learning data for classification; and

cause a model to learn classification of the feature extracted from the log data using the feature vector of the failure log to which the inverse vector of the feature vector of the normal system log is added.

7. An information providing method performed by a computer and comprising:

acquiring log data generated by an information equipment system;

calculating a first term frequency-inverse document frequency (TF-IDF) vector to extract a feature of the log data;

calculating a second TF-IDF vector as a feature of information included in history information related to the information equipment system;

acquiring information from the history information based on a degree of similarity between the first TF-IDF vector and the second TF-IDF vector;

outputting the acquired information;

generating a model for calculating likelihood of classifying a log for each type of failure and for each time interval acquired by dividing a time until a failure occurs; and

detecting a sign of failure occurrence of the information equipment system by using the model and the extracted feature of the log data.

Assignments (2)

Priority Claims (1)

Continuity (1)

References Cited (11)