Design And Implementation Of Information System Diagnosis Platform Based On Time Series Mining

In recent years,with the continuous development of social production methods and informatization levels,the operation and maintenance of information systems have gradually attracted people's attention.In the context of the widespread application of artificial intelligence,how to use big data technology to improve the intelligence level in the operation and maintenance field,and reduce the pressure on operation and maintenance,which has become the primary problem that needs to be broken.This thesis designs and implements an information system diagnostic platform based on time-series mining.First,it mainly proactively discovers potential failures of information systems by performing time series analysis and abnormal mining of complex business indicator data.Then,it builds an evaluation system of the operating status of the information system to achieve real-time monitoring of the "health status" of the information system.Finally,it also combined with the improved prediction algorithm to complete the prediction of the overall situation of the system in order to find early signs and laws of abnormal operation of information systems.Synthesizing the "diagnosis-evaluation-prediction" three-phase operation and maintenance plan proposed by the process,the main research contents of this thesis are as follows:1.Research on anomaly detection algorithms based on unified timing analysis framework.On the basis of traditional timing monitoring,this thesis builds a unified timing analysis framework that combines the integrates batch processing and modeling of offline large-scale time series data with the calculation and anomaly detection of online real-time streaming data,which greatly improves the hit rate of anomaly detection in actual business systems,and has provided effective data support for subsequent information system status assessment and trend prediction.2.Research on system health assessment model based on time series analysis optimization.By comprehensively comparing the differences between the application of traditional assessment methods,this thesis refines the influence weight of indicator based on fuzzy matrix,and weights the horizontal and vertical relationships of each indicators during the construction of the core judgment matrix through time series mining technology.Then,it also optimizes the consistency in combination with genetic algorithms to guarantees the rationality and effectiveness of the model and achieve amore accurate evaluation of the operating status of information systems.3.Research on the system situation prediction model based on hybrid neural network.Considering that the prediction of the system situation is a combination of long-term and short-term analysis,this thesis based on the traditional short-term time series prediction and combined with TreNet hybrid neural network for long-term trend analysis,proposed an improved situation prediction model,which has broken the bottleneck of traditional timing prediction that is difficult to apply in actual business scenarios and has low accuracy.4.Design and implement an information system diagnostic platform based on time series mining.Based on the operation and maintenance requirements of the actual information system,this thesis analyzes and designs the overall functional architecture of the diagnostic platform,and uses Html5 design to implement the graphical user interface.The core algorithm of the system is implemented based on Python and uses MySql as a database to store timing information and results.This thesis tests the system functions and algorithms,and verifies the effectiveness and usability of the information system diagnostic platform.Which has positive significance for the development of intelligent operation and maintenance in the context of big data.