System Introduction
Bigdata Analysis system integrates with real-time information system and other solution’s data to analyze large amount of data through multivariate statistical analysis techniques such as PLS (Partial Least Squares) or PCA (Principal Component Analysis).
Main functions
PCA(Principal Component Analysis)
- The model is derived by analyzing the principal component statistical information among variables using history data.
- The data from large number of variables is reduced in dimensionality, minimizing information loss while simplifying and making it easier to understand the complex structure among correlated variables.
- It is good to monitor which variables are showing different data pattern(anomaly) while other variables are normal.
PLS(Projection to Latent Structure)
- The model is derived not only considering the correlations among explanatory variables (x) based on past data but also taking into account the correlations of the response variable (y).
- PLS (Partial Least Squares) not only derives predictions for the response variable to detect abnormal states in the process but also considers past relationships to generate predicted values even in areas where data measurement is not available, enabling process optimization.
Application Field
PLS and PCA Statistical Models
- Soft Sensors
- Energy Models
- Process Models
Use Case
