T-ray relevant frequencies for osteosarcoma classification

Abstract

We investigate the classification of the T-ray response of normal human bone cells and human osteosarcoma cells, grown in culture. Given the magnitude and phase responses within a reliable spectral range as features for input vectors, a trained support vector machine can correctly classify the two cell types to some extent. Performance of the support vector machine is deteriorated by the curse of dimensionality, resulting from the comparatively large number of features in the input vectors. Feature subset selection methods are used to select only an optimal number of relevant features for inputs. As a result, an improvement in generalization performance is attainable, and the selected frequencies can be used for further describing different mechanisms of the cells, responding to T-rays. We demonstrate a consistent classification accuracy of 89.6%, while the only one fifth of the original features are retained in the data set.

Support Vector Machine Applications in Terahertz Pulsed Signals Feature Sets

“…First, the structural risk minimization problem is solved to reveal, among all the learning vectors, those that have the minimum margin to the optimal decision hyperplane, these vectors are called the support vectors. Finally, the decision hyperplane is defined by the following equation: (4) where , are support vectors. The solution of this large-scale quadratic programming problem is applied to calculate the coefficients and .…”

“…(i) The carrier substrate is a commercially available cyclic olefin homopolymer slide (Greiner Bio-One) with transparent and dispersion-free properties in the T-ray frequency range. (ii) In order to achieve T-ray imaging, small liquid volumes are spotted by hand in a 4 4 array of alternating poly-A and poly-C on the slide. Each spot with a diameter of approximately 1 mm consists of 2 of deionized water containing 0.2 mg material.…”

“…This obviously reduces the computational resource requirements, which is attractive for possible hardware-based implementations. Further, the sparse features help us avoid the overfitting problem [4], [23]. The input vectors occupy a size matrix, where is equal to the number of input vectors (training vectors) and is the dimensionality of each feature vector.…”

“…T-rays have promising potential both in in vivo and in vitro biosensing applications [1]- [3] owing to: 1) their noninvasive property and 2) the fact that biomolecules have rich resonances in the T-ray region [4]- [6]. The ultimate aim of our work is to perform automatic classification of data obtained from T-ray measurements with tomographic applications [7].…”

Fiber-based Terahertz Time-Domain Spectroscopy Systems Operated in the Telecom Band

“…Although, algorithms that are more accurate have been developed, taking into account signal echoes (Fabry-Pérot-oscillations) [111,112,113,114,115,116] or providing increasing accuracy with an iterative method [115].…”