A comparative analysis of classification methods for hyperspectral images generated with conventional dimension reduction methods

Abstract

This paper compared performances of classification methods for a hyperspectral image dataset in view of dimensionality reduction (DR). Among conventional DR methods, principal component analysis, maximum noise fraction, and independent component analysis were used for the purpose of dimension reduction. The study was conducted using these DR techniques on a real hyperspectral image, an AVIRIS dataset with 224 bands, throughout the experiments. It was observed that DR may have a significant effect on the classification performance. After the DR methods were applied to the image dataset, the extracted reduced bands were used for testing classification performances. Four commonly used classification methods including maximum likelihood, Mahalanobis distance, spectral angle mapper, and support vector machines were used for the classification of dimensionally reduced and original (not reduced) images to test whether classification accuracies differed significantly for these images. Experiments reported in this study indicate that second-order statistics-based DR methods could efficiently reduce the dimensionality of the hyperspectral dataset.The maximum likelihood method generally performed better than the other classifiers while the spectral angle mapper exhibited significantly lower accuracies in the experiments. Inclusion of a texture measure in classification improved the classification results in a significant manner.

Comparison and Evaluation of Dimensionality Reduction Techniques for Hyperspectral Data Analysis

“…The scree test is a criterion used as an optimal method for reducing the dimensionality band from the hyperspectral dataset [6]. From the scree plot, it is shown that the first nine principal components are retained and the bands corresponding to these components provide large data variance when compared with the low-order PC.…”

“…Classification results are greatly influenced by MNF when the ground cover features exhibit homogeneity. The efficiency of the classifiers improves with reduced components that reveal apparently informative bands [5,6]. This paper compares and evaluates the performance of two defined DR methods, namely PCA and MNF, by interpreting the eigen values acquired during processing.…”

Comparison of Statistical Underlying Systematic Risk Factors and Betas Driving Returns on Equities

“…In this context, in a first study, Ladrón de Guevara & Torra (2014) estimated the underlying structure of systematic risk by using Principal Component Analysis (PCA) and Factor Analysis (FA) 2 ; it included the testing of the models in two versions: returns and returns over the riskless interest rate for weekly and daily databases, and a two-stage methodology for the econometric contrast. First, they extracted the underlying systematic risk factors using both the standard linear version of Principal Component Analysis and the maximum likelihood Factor Analysis estimation, and they were able to reconstruct the observed returns using the factors extracted almost perfectly in all cases.…”