Robustness of Partial Least-Squares Models to Change in Sample Temperature: II. Application to Fruit Attributes

Abstract

Partial least-squares regression models were developed using spectra of tomato fruit collected at 15°C and tested on spectra of an independent set of fruit at higher sample temperatures. The influence of sample temperature on the model used to predict fruit dry matter (DM) was manifested primarily in terms of bias, not standard error of prediction. For example, a model for DM created with samples at 15°C had a bias of-0.9% DM and-1.9% DM when used to predict DM in fruit at 25°C and 35°C, respectively. The addition of spectra of a relatively small number of samples collected at different temperatures to the calibration set can create a model that is robust to temperature; however, continued addition of samples at a uniform temperature overwhelms this compensation effect, resulting in a model that is not robust to temperature. For a model that included spectra of fruit at a range of temperatures, the prediction bias increased as the ratio of samples at 15°C to samples at other temperatures increased beyond 200:1. The use of orthogonal scatter correction, external parameter orthogonalisation, generalised least-square weighting, global model development and repeatability file were compared for the development of a temperature-robust DM model. The use of a repeatability file is recommended on the basis of the lowest root mean square error of prediction and bias. Selection of wavelength regions to avoid water absorption features is recommended for an attribute not associated with an OH feature (such as skin-colour prediction).

Factors influencing near infrared spectroscopy analysis of agro-products: a review

“…Additionally, the global model showed useful prediction ability, which made temperature fluctuation a negligible interference. Besides, there are many other studies [37][38][39][40][41][42][43] that have proved that the global model can significantly reduce the effect of temperature on NIR measurement. However, the drawback of this approach is that the required data collection is quite a large undertaking because the local temperature model uses a calibration set for a single selected temperature and a validation set for other temperatures, while the global temperature model contains all of the data sets.…”

“…However, the drawback of this approach is that the required data collection is quite a large undertaking because the local temperature model uses a calibration set for a single selected temperature and a validation set for other temperatures, while the global temperature model contains all of the data sets. Based on large scale of data sets, Acharya et al [41] designed several methods of population structuring for detecting the dry matter and color of tomatoes with the aim of producing robust models that take account of sample temperature. It was concluded that temperature compensation created by adding spectra of the same set of samples at different temperatures was overwhelmed by continuing addition of 500 spectra at a uniform temperature, resulting in a model that was not robust to temperature.…”

NIR spectroscopy in Nepal

“…However human capacity is developing, e.g., Dr. Umesh Kumar Acharya of NARC obtained a PhD on the use of NIRS in the assessment of fruit quality at Central Queensland University, Australia. 3,4…”

Assessment of internal flesh browning in intact apple using visible-short wave near infrared spectroscopy

“…Fruit consist of approximately 80% water. The IR/NIR spectrum of water is affected by temperature due to an effect on the extent of hydrogen bonding, and this can impact prediction of soluble solids content (SSC) and dry matter (DM) in intact fruit using SWNIRS with the simplest accommodating measure being the inclusion of samples of a range of temperatures into the training sets (Acharya et al, 2014). Given that the spectral information relevant to internal browning may be restricted to the visible region, spectral based measures of internal browning may be free from interference associated with temperature change, however, this issue should be explicitly considered.…”