Evaluation of machine learning methods for organic apple authentication based on diffraction grating and image processing

Optical measuring technologies coupled with machine learning algorithms can be used to build a home-made sensor system. We built such a sensor system using a smartphone and a diffraction grating sheet. Diffraction images were captured under white light illumination and converted into a data matrix for data analysis. In this paper we present a systematic evaluation of this sensor system on the task of differentiating organic apples from conventional ones. We used the sensor system to measure 150 organic and conventional apples as rainbow images. We processed the rainbow images using computer vision techniques, built machine learning and chemometrics models, and used the resultant models to classify testing samples. Moreover, a comparative study was conducted where the same set of apples were scanned by a commercial spectrometer resulting in spectral data of the apple samples and classification was undertaken using partial least squares discriminant analysis (PLS-DA). Experimental results show that state of the art machine learning algorithms such as support vector machine (SVM) and locally weighted partial least squares classifier (LW-PLSC) are effective in handling low-quality image data with classification accuracies of 93 − 100%. These results suggest that the sensor system is convenient and low-cost, and provides a fast, effective, non-destructive and viable solution for in-line food authentication.     

Early prediction of lung cancer based on the combination of trace element analysis in urine and an Adaboost algorithm

Early detection of cancer is the key to effective treatment and long-term survival. Lung cancer is one of the most frequently occurring cancers and its early detection is particularly of interest. This work investigates the feasibility of a combination of Adaboost (ensemble from machining learning) using decision stumps as weak classifier and trace element analysis for predicting early lung cancer. A dataset involving the determination of 9 trace elements of 122 urine samples is used for illustration. Kennard and Stone (KS) algorithm coupled with an alternate re-sampling was used to realize sample set partitioning. The whole dataset was split into equally sized training and test set, which were then reversed to yield a second operating case, we called them case A and case B, respectively. The prediction results based on the Adaboost were compared with those from Fisher discriminant analysis (FDA). On the test set, the final Adaboost classifiers achieved a sensitivity of 100% for both cases, a specificity of 93.8%, 95.7%, and an overall accuracy of 95.1%, 96.7%, for case A and case B, respectively. In either case, Adaboost always achieves better performance than FDA; also, it is less sensitive to the composition of the training set compared to FDA and easy to control over-fitting. It seems that Adaboost is superior to FDA in the present task, indicating that integrating Adaboost and trace element analysis of urine can serve as a useful tool for diagnosing early lung cancer in clinical practice.     

Monitoring of High-shear Granulation using Acoustic Emission: Predicting Granule Properties

Sound in the ultrasonic (20 to 1,000 kHz) range emitted during high-shear granulation was recorded and analyzed using multivariate techniques in order to assess the relationship between variations in the physical properties of the obtained granules and the evolution of acoustic emissions taking place during their formation. The acoustic signal analysis was preformed on two different granulators. A four-component model was obtained from the analysis and was capable of predicting the particle size distribution of the granules based on a 13 sieve cut measurement covering the range of 53–1,180 μm. The average error of prediction was less than 2%. Acoustic emission also proved useful as a predictor of granule moisture content. The final granule moisture content could be predicted with a root-mean-square error of prediction of 5.76% and 1.9% in the two different granulators evaluated in this investigation. The acoustic signals emitted during wet granulation contain the information necessary to make quantitative assessments of the changes in water content, particle size, and the particle size distribution of the granules produced by the process.     

化学计量学在中药定量表征中的应用

随着中药现代化的不断深入,对中药定量化、标准化,即对中药的定量表征,已日益得到研究人员的广泛认同。而化学计量学通过应用数学、统计学和计算机技术的原理和方法来处理中药的复杂数据,并将多变量的研究引入其中从而良好地对中药进行定量表征,更是取得了较大的进展。该文对近年来中药定量表征中存在的问题、主要化学计量学方法的原理、特点、局限性及常用统计方法以及其在中药定量表征中的应用现状进行综述,以期对深入研究中药定量表征提供参考,为多种化学计量学方法的结合使用提供进一步的研究思路。     

A comparison of the use of refractive index (RI) and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) for the provenance establishment of glass bottles

The use of laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) has been compared with the traditional method of refractive index (RI) measurement for the establishment of the provenance of glass bottles. Using the RI method alone, it is not possible to discriminate between certain glass bottles produced up to 18 days apart from a single manufacturing plant. Furthermore, variations in RI within a single bottle can be large enough to invalidate co-provenance establishment using this technique alone. Determination of the trace elemental composition of bottles collected over a 1-month period confirmed that minimal variation of trace metal distribution occurred within individual bottles made during this period. Therefore, the trace element composition of any fragment of glass from a broken bottle may be considered representative of the elemental composition of the entire bottle. In addition, statistical comparison of the distribution of approximately 38 of the 56 analytes that were determined established that it was possible to discriminate between two glass bottles manufactured in the same plant two hours apart. Using this methodology it has been possible to develop an analytical protocol to significantly improve the accurate comparison and provenance establishment of forensic glass evidence.     

Comparison of some secondary metabolite content in the seventeen species of the Boraginaceae family

Context: The Boraginaceae family comprises plants that have important therapeutic and cosmetic applications. Their pharmacological effect is related to the presence of naphthaquinones, flavonoids, terpenoids, phenols, or purine derivative – allantoin.

Objective: In the present study, comparison of some secondary metabolite content and phytochemical relationship between 17 species of the Boraginaceae family were analyzed.

Materials and methods: High performance capillary electrophoresis (HPCE) was used to perform a chemometric analysis in the following Boraginaceae species: Anchusa azurea Mill., Anchusa undulata L., Borago officinalis L., Buglossoides purpurocaerulea (L.) I.M. Johnst., Cerinthe minor L., Cynoglossum creticum Mill, Echium italicum L., Echium russicum J.F. Gmel., Echium vulgare L., Lindelofia macrostyla (Bunge) Popov (syn. Lindelofia anchusoides (Lindl.) Lehm.), Lithospermum officinale L., Nonea lutea (Desr.) DC., Omphalodes verna Moench (syn. Cynoglossum omphaloides L.), Pulmonaria mollis Wulfen ex Hornem., Pulmonaria obscura Dumort., Symphytum cordatum Waldst. & Kit ex Willd., and Symphytum officinale L.

Results: Six active compounds in shoot extracts (allantoin, p-hydroxybenzoic acid, rutin, hydrocaffeic acid, rosmarinic acid, and chlorogenic acid) and four compounds in root extracts (allantoin, hydrocaffeic acid, rosmarinic acid, and shikonin) were identified. The presence and abundance of these compounds were used for the characterization of the species and for revealing their phytochemical similarity and differentiation.

Discussion and conclusion: The present study provides the first comprehensive report of the extraction and quantification of several compounds in Boraginaceae species (some of them for the first time). Among the 17 species studied, species with potentially high pharmacological activity were recognized.     

Spectrophotometric monitoring of nimesulide photodegradation by a combined hard-soft multivariate curve resolution-alternative least square method

Nimesulide, a Cox-2 inhibitor anti-inflammatory drug, is a light sensitive compound and its biological activity is decreased upon photodegradation. The photodegradation kinetic of nimesulide was investigated spectrophotometrically using multivariate curve resolution analysis to overcome spectral overlapping of reactant and degradation products. The absorbance spectra of the nimesulide methanolic solution, exposed to daily sunlight, were recorded at different times. Three absorbing chemical species, coexisted in the reaction system, were detected by application of factor analysis to the absorbance data matrix. The soft-modeling multivariate curve resolution-alternative least square (MCR-ALS) analysis of the evolutionary absorbance data revealed that nimesulide undergoes photodegradation through a two-step consecutive manner, where both steps obey first-order kinetic. By application of the kinetic hard model constraint to the MCR-ALS analysis, an excellent agreement was obtained between the fitted concentration profiles and those obtained by soft method. The first-order rate constants of the first and second degradation products were calculated as 0.052 (+/-0.007) and 0.009 (+/-0.001)h(-1), respectively. Finally, the pure spectra of the resolved chemical species were calculated, where that of nimesulide was the same as that obtained experimentally.     

Isotopic ratios to detect infringements of patents or proprietary processes of pharmaceuticals: two case studies

Because of the increasing problem of drug counterfeiting and the potential danger related as well as the economic losses involved, the pharmaceutical industry and the regulatory instances are interested in the development of anti-counterfeiting and patent protection methodologies. In this paper, the evaluation of measured isotopic ratios by means of explorative chemometric techniques was performed to distinguish groups in two data sets containing samples of acetyl salicylic acid and ibuprofen, respectively. The samples in the data sets originated from different countries and manufacturers. For both compounds a clear distinction of groups of samples could be obtained. These groups could be explained based on the origin of the samples, both geographically as well as based on the manufacturer. Hypotheses were formulated concerning the synthetic pathways of the molecules and they were linked to the groups obtained with the chemometric tools.     

Figures of Merit Comparison of Reflectance and Transmittance Near-Infrared Methods for the Prediction of Constituent Concentrations in Pharmaceutical Compacts

The purpose of this work was to demonstrate how multivariate figures of merit, in conjunction with accuracy and precision statistics, can be used to evaluate the performance of near-infrared reflectance and transmittance spectroscopy for the nondestructive chemical quantification of intact pharmaceutical compacts. A total of 174 four-component compacts were produced at five compaction pressures following a fully balanced quaternary design. Near-infrared spectra were measured in both reflectance and transmittance modes. Partial least-squares regression was used to model spectroscopic response, and net analyte signal theory was used to generate the figures of merit. It was discovered that sample positioning variation more negatively impacted the performance (analytical sensitivity, S/N ratio, limit of detection) of the transmittance method. Process analytical technology implementation teams should consider these aspects during technology selection, method development, and validation.