Fast Classification of Brick Samples by Combination of Principal Component Analysis and Linear Discriminant Analysis Using Standoff Laser-Induced Breakdown Spectroscopy
| Authors | |
|---|---|
| Year of publication | 2013 |
| Type | Conference abstract |
| MU Faculty or unit | |
| Citation | |
| Description | LDA is a multivariate statistical method for discrimination of objects up to a finite number of categories, based on a certain subset of all objects (training set). The principle of this method is maximizing the ratio of the between-class variance to the within-class variance. The decision rules obtained by classification of training set are later applied to the testing set. PCA was applied to reduce a dimensionality and the scores were computed from whole spectra (200 – 1000 nm) of 29 brick samples obtained by stand-off LIBS. After dividing the sample set into training and test samples first few principal components were used as inputs to LDA and the predictions of localities were computed. The results suggest that stand-off LIBS in combination with advanced statistical methods has a big potential for archaeological in-field measurements. By examining of PCA scores there was also discovered a dependence of firing temperature of bricks. Spectra of bricks fired on 300 – 1000 Celsius differs and this property is reflected significantly in only one of the principal components. |
| Related projects: |