- 专利标题: Calibratable log projection and error remediation system
-
申请号: US16449644申请日: 2019-06-24
-
公开(公告)号: US10713143B1公开(公告)日: 2020-07-14
- 发明人: Jayanti Vemulapati , Lisa Suzanne Wilson , Murtuza Chitalwala , Chethana Savalgi , Sekhar A. Badugu
- 申请人: Accenture Global Solutions Limited
- 申请人地址: IE Dublin
- 专利权人: Accenture Global Solutions Limited
- 当前专利权人: Accenture Global Solutions Limited
- 当前专利权人地址: IE Dublin
- 代理机构: Brinks Gilson & Lione
- 主分类号: G06F11/34
- IPC分类号: G06F11/34 ; G06N20/00 ; G06F9/50 ; G06F9/455
摘要:
A system access a session profile. The session profile may include log source identifiers and model identifiers. The system may deploy a log projection session based on the session profile. The system may receive, in response to deployment of the log projection session, a log stream from a log source corresponding to at least one of the log identifiers. The system may generate a log projection stream based on the log stream and an initial machine-learning model. The system may calibrate the session profile and select an alternative machine-learning model based on model performance metrics. The system may redeploy the log projection session based on the calibrated session profile. The system may automatically scale computer resources for improved job performance based on forecasted log information derived from the selected machine-learning model.
信息查询