Partial Least-squares Support Vector Machine Research Articles

Low-abundance proteins-based label-free SERS approach for high precision detection of liver cancer with different stages

BackgroundSurface-enhanced Raman scattering (SERS) technology have unique advantages of rapid, simple, and highly sensitive in the detection of serum, it can be used for the detection of liver cancer. However, some protein biomarkers in body fluids are often present at ultra-low concentrations and severely interfered with by the high-abundance proteins (HAPs), which will affect the detection of specificity and accuracy in cancer screening based on the SERS immunoassay. Clearly, there is a need for an unlabeled SERS method based on low abundance proteins, which is rapid, noninvasive, and capable of high precision detection and screening of liver cancer. ResultsSerum samples were collected from 60 patients with liver cancer (27 patients with stage T1 and T2 liver cancer, 33 patients with stage T3 and T4 liver cancer) and 40 healthy volunteers. Herein, immunoglobulin and albumin were separated by immune sorption and Cohn ethanol fractionation. Then, the low abundance protein (LAPs) was enriched, and high-quality SERS spectral signals were detected and obtained. Finally, combined with the principal component analysis-linear discriminant analysis (PCA-LDA) algorithm, the SERS spectrum of early liver cancer (T1-T2) and advanced liver cancer (T3-T4) could be well distinguished from normal people, and the accuracy rate was 98.5% and 100%, respectively. Moreover, SERS technology based on serum LAPs extraction combined with the partial least square-support vector machine (PLS-SVM) successfully realized the classification and prediction of normal volunteers and liver cancer patients with different tumor (T) stages, and the diagnostic accuracy of PLS-SVM reached 87.5% in the unknown testing set. SignificanceThe experimental results show that the serum LAPs SERS detection combined with multivariate statistical algorithms can be used for effectively distinguishing liver cancer patients from healthy volunteers, and even achieved the screening of early liver cancer with high accuracy (T1 and T2 stage). These results showed that serum LAPs SERS detection combined with a multivariate statistical diagnostic algorithm has certain application potential in early cancer screening.

Identification of Healthy Tissue from Malignant Tissue in Surgical Margin Using Raman Spectroscopy in Oral Cancer Surgeries.

Accurate identification of tissue types in surgical margins is essential for ensuring the complete removal of cancerous cells and minimizing the risk of recurrence. The objective of this study was to explore the clinical utility of Raman spectroscopy for the detection of oral squamous cell carcinoma (OSCC) in both tumor and healthy tissues obtained from surgical resection specimens during surgery. This study enrolled a total of 64 patients diagnosed with OSCC. Among the participants, approximately 50% of the cases were classified as the most advanced stage, referred to as T4. Raman experiments were conducted on cryopreserved tissue samples collected from patients diagnosed with OSCC. Prominent spectral regions containing key oral biomarkers were analyzed using the partial least squares-support vector machine (PLS-SVM) method, which is a powerful multivariate analysis technique for discriminant analysis. This approach effectively differentiated OSCC tissue from non-OSCC tissue, achieving a sensitivity of 95.7% and a specificity of 93.3% with 94.7% accuracy. In the current study, Raman analysis of fresh tissue samples showed that OSCC tissues contained significantly higher levels of nucleic acids, proteins, and several amino acids compared to the adjacent healthy tissues. In addition to differentiating between OSCC and non-OSCC tissues, we have also explored the potential of Raman spectroscopy in classifying different stages of OSCC. Specifically, we have investigated the classification of T1, T2, T3, and T4 stages based on their Raman spectra. These findings emphasize the importance of considering both stage and subsite factors in the application of Raman spectroscopy for OSCC analysis. Future work will focus on expanding our tissue sample collection to better comprehend how different subsites influence the Raman spectra of OSCC at various stages, aiming to improve diagnostic accuracy and aid in identifying tumor-free margins during surgical interventions.

Label-free detection of bladder cancer and kidney cancer plasma based on SERS and multivariate statistical algorithm

In this study, we mainly aimed to investigate the diagnostic potential of surface-enhanced Raman spectroscopy for bladder cancer and kidney cancer which are the most common cancers of the urinary system, and evaluate the classification ability of three statistical algorithms: principal component analysis-linear discriminate analysis (PCA-LDA), partial least square-random forest (PLS-RF), and partial least square-support vector machine (PLS-SVM). The plasma of 26 bladder cancer patients, 38 kidney cancer patients and 39 normal subjects was mixed with the same volume of silver nanoparticles, respectively, and then high-quality SERS signal was obtained. The SERS spectra in the range of 400–1800 cm−1 were compared and analyzed. There were some significant differences in SERS peak intensity, which may reflect the changes in the content of some biomacromolecules in the plasma of cancer patients. Based on the three algorithms of PCA-LDA, PLS-RF and PLS-SVM, the classification accuracy of SERS spectra of plasma from cancer patients and normal subjects was 98.1%, 100% and 100%, respectively.In addition, the classification accuracy of the three diagnostic algorithms to classify the SERS spectra of bladder cancer and kidney cancer was 81.3%, 91.7%, and 98.4%, respectively. This exploratory work demonstrates that SERS combined with PLS-SVM algorithm has superior performance for clinical screening of bladder cancer and kidney cancer through peripheral plasma.

Nondestructive Prediction of Rice Seed Viability Using Spectral and Spatial Information Modeling of Visible–Near Infrared Hyperspectral Images

HighlightsAn NIR-Vis hyperspectral imaging approach was developed to predict the viability of rice seeds.Through multi-step accelerated aging, seed lots in various states were used for the experiments.Models using spectral information and spectral-spatial information of hyperspectral images were used and compared.Abstract. Rice is one of the world’s most important food crops, and rice seed viability is an important factor in rice crop production. In this study, a visible–near infrared (vis–NIR) hyperspectral imaging system and spectral–spatial information modeling are used to predict the viability of rice seeds. Experimental samples are prepared using seeds harvested in two different years and artificially aged for various periods. Vis-NIR hyperspectral acquisition and germination tests of the prepared seed samples are performed. Partial least square (PLS)–discriminant analysis, a support vector machine (SVM), a PLS–SVM, a PLS–artificial neural network, and a one-dimensional–convolutional neural network (CNN) for the mean spectra of seeds, as well as a CNN, a PLS–CNN, and dual branch networks for the hyperspectral images of the seeds are applied for viability prediction modeling. Result shows that an accuracy of approximately 90% and high f1 scores can be obtained in most models. Furthermore, it is confirmed that models using spectral and spatial information can classify hard samples more effectively. Keywords: Deep learning, Hyperspectral images, Rice, Seed, Spectroscopy, Viability.

Quantification of multiple soil trace elements by combining portable X-ray fluorescence and reflectance spectroscopy

Several combinations of proximal sensors, portable x-ray fluorescence spectroscopy (pXRF), visible-near-infrared spectroscopy (vis-NIR), and mid-infrared spectroscopy (MIR) are assessed to quantify multiple soil trace elements (TE) with reliable accuracy. A total 622 topsoil samples (0–20 cm depth) collected at regular 8 km spacing across southern New Zealand were analysed for TE concentrations by a reference laboratory method and scanned with pXRF, vis-NIR, and MIR. The laboratory results and spectral information were used to develop robust models that enable prediction for soil concentrations of arsenic (As), cadmium (Cd), chromium (Cr), copper (Cu), mercury (Hg), nickel (Ni), lead (Pb), and zinc (Zn). Chemometric approaches including spectral data fusion, partial least squares – support vector machine (PLS-SVM) model fusion, and model averaging were tested to obtain optimal quantitative TE predictions in the samples. Validation of PLS-SVM based models on held-out samples showed the combination of pXRF and MIR performed optimally for predicting As, Cr, and Pb concentrations with concordance correlation coefficient (CCC) of 0.97 and root mean square error (RMSE) of 0.57 mg As/kg, 1.14 mg Cr/kg, and 1.05 mg Pb/kg respectively. Fusion of pXRF and vis-NIR data performed optimally for quantifying Cu, Ni, and Zn concentrations with CCC of 0.98 and RMSE of 1.09 mg Cu/kg, 0.63 mg Ni/kg, and 3.90 mg Zn/kg respectively. For Cd and Hg, Granger-Ramanathan model averaging of outputs from PLS-SVM based pXRF, vis-NIR, and MIR models performed optimally with CCC > 0.80 and RMSE of 0.03 mg Cd/kg and 0.01 mg Hg/kg respectively. This study showed that quantitative prediction models for As, Cd, Cr, Cu, Hg, Ni, Pb, and Zn can be successfully implemented at local- and national-scale and for long-term monitoring of soil trace elements at concentrations below pXRF detection limits and with reduced matrix interference from organic matter than from individual techniques alone.

Classification and Quantitative Modeling Analysis of Groundwater Hardness based on Ultraviolet Absorption Spectrum

In accordance with the national standard of groundwater quality, samples with various ion concentrations are prepared based on water hardness standard substances, and a water hardness classification and quantitative analysis model based on ultraviolet absorption spectrum are established. Firstly, thepartial least squares discriminant analysis (PLSDA) model is established by using a group of samples with different ion concentration that form the training set, and the binary tree is introduced for multiple classification of water hardness. Secondly, partial least squares support vector machine (PLS-SVM) is used to establish the quantitative model of each class of samples. Atlast, Monte carlo cross-examination is performed on the model using a test set composed of samples with different ion concentrations. Experimental results show that, for the identification of underwater hardness, the average error rate is 1.13%, which is less than other common classification methods, and MSE is 0.247, and R2 is 0.987, which both are improved than other common methods for the prediction of ion concentration of underwater.

Strategies for the efficient estimation of soil organic matter in salt-affected soils through Vis-NIR spectroscopy: Optimal band combination algorithm and spectral degradation

Visible and near-infrared (Vis-NIR) spectroscopy is a cost-effective technique for alternative soil physical and chemical analyses for estimating soil properties. The optimal band combination algorithm is an effective method of extracting spectral variables by considering the interaction information between wavebands, but for laboratory Vis-NIR spectral data, this method is susceptible to the “dimensional curse”. Here, we hypothesized that properly degrading the spectral configuration (i.e., decreasing the number of spectral bands and coarsening the spectral resolution) can improve the computational efficiency without affecting the prediction accuracy. To test this hypothesis, we constructed six degraded spectral configurations from an initial spectral database (i.e., consisting of 2001 spectral bands acquired with a portable ASD spectroradiometer) with a reduction in the number of spectral bands from 2001 to 19, a coarsened spectral resolution from 3 to 100 nm, and a spectral sampling interval equal to the spectral resolution (i.e., uniform interval sampling). In this study, the databases consisted of 255 soil samples collected from the Ebinur Lake area in Northwest China. The relationship between the soil organic matter (SOM) and the spectra was established using a partial least-squares-support vector machine (PLS-SVM) through two strategies: one is in accordance with the different salinity levels, and the other involves applying the optimal band combination algorithm from each spectral configuration. The results indicated that the soil salinity had a strong negative influence on the performance of the SOM models (R2cv, 0.46–0.81). However, the optimal band combination algorithm can improve the sensitivity (R2pre, 0.36–0.65) of spectral information and the SOM. Overall, the prediction accuracy obtained through the optimal band combination algorithm was generally superior to that from full-spectrum data. The prediction performance of the optimal band combination algorithm was accurate (R2pre ≥ 0.85) and stable (RPIQ pre, ~3.20), with a spectral resolution between 3 and 20 nm (i.e., the number of spectral bands decreased from 2001 to 99). Considering the accuracy and time-consuming nature of this approach, the combination of a 20 nm spectral resolution and an optimal band combination algorithm was the most effective method. In summary, this research will guide future studies in transforming hyperspectral datasets into parsimonious representations and uses the optimal band combination algorithm efficiently to determine the informative variable. Furthermore, the optimal band combination algorithm has broad application prospects in soil Vis-NIR spectroscopy and other fields of spectroscopy.

Prediction of soil organic matter in northwestern China using fractional-order derivative spectroscopy and modified normalized difference indices

Visible-near-infrared (Vis-NIR) spectroscopy makes it possible to estimate soil organic matter content (SOMC). Spectral pretreatment techniques have important significance in the quantitative analysis of SOMC. A total of 150 soil samples collected in northwestern China were used as data sets for calibration and validation. The SOMC values and reflectance spectra were measured in the laboratory. Fractional-order derivatives (FODs) (intervals of 0.05, range of 0–2) were used for soil spectral pretreatment, and a new three-band index (modified normalized difference index, MNDI) was constructed based on the band-optimization algorithm and the existing two-band exponential form (normalized difference index, NDI). Partial least square-support vector machine (PLS-SVM) models were calibrated using spectral parameters selected based on a single dimension (FOD), two-dimensional index (NDI) and three-dimensional index (MNDI) and subsequently applied to estimate SOMC. Three model evaluation parameters, namely, the coefficient of determination (R2), root mean squared error (RMSE), and ratio of performance to interquartile range (RPIQ), were used to evaluate the estimation accuracy of the models. The results showed that with increased derivative order, the spectral strength gradually decreased, but the spectral detail increased. Furthermore, the correlation between FOD spectra and SOMC was significantly enhanced in the visible region, with the most obvious effect in the 1.05- to 1.45-order range. The PLS-SVM modeling results showed that the sensitivity and estimation accuracy of SOMC increased with increasing spectral synergy (i.e., 1D (FOD) < 2D (NDI) < 3D (MNDI)). Among the models, MNDI exhibited the best model performance, yielding a validation R2 and RPIQ of 0.846 and 3.396, respectively. The combination of FOD and MNDI could weaken the soil noise information and improve the prediction accuracy of SOMC. Furthermore, the three-dimensional index has strong application potential for estimating other biochemical parameters of soil using Vis-NIR spectroscopy.

Edible Gelatin Diagnosis Using Laser-Induced Breakdown Spectroscopy and Partial Least Square Assisted Support Vector Machine.

Edible gelatin has been widely used as a food additive in the food industry, and illegal adulteration with industrial gelatin will cause serious harm to human health. The present work used laser-induced breakdown spectroscopy (LIBS) coupled with the partial least square–support vector machine (PLS-SVM) method for the fast and accurate estimation of edible gelatin adulteration. Gelatin samples with 11 different adulteration ratios were prepared by mixing pure edible gelatin with industrial gelatin, and the LIBS spectra were recorded to analyze their elemental composition differences. The PLS, SVM, and PLS-SVM models were separately built for the prediction of gelatin adulteration ratios, and the hybrid PLS-SVM model yielded a better performance than only the PLS and SVM models. Besides, four different variable selection methods, including competitive adaptive reweighted sampling (CARS), Monte Carlo uninformative variable elimination (MC-UVE), random frog (RF), and principal component analysis (PCA), were adopted to combine with the SVM model for comparative study; the results further demonstrated that the PLS-SVM model was superior to the other SVM models. This study reveals that the hybrid PLS-SVM model, with the advantages of low computational time and high prediction accuracy, can be employed as a preferred method for the accurate estimation of edible gelatin adulteration.

Nonlinear Multivariate Quality Prediction Based on OSC-SVM-PLS

Predicting data with noise and high correlation using latent structural projection or partial least-squares (PLS) is a commonly used linear regression method. In order to solve the nonlinear problem existing in the industrial processes, many methods for extending the PLS to the nonlinear space have been proposed. In this paper, the support vector machine (SVM) is integrated into the PLS to get a nonlinear PLS model. In order to reduce information in process variables that are independent of quality variables, a orthogonal signal correction (OSC) method has been introduced. Nonlinear data from industrial processes is used to build nonlinear PLS models and SVM methods are adopted to approximate nonlinear internal relationships between the input and the output. The model proposed in this paper can predict industrial process quality variables and improve prediction accuracy compared with traditional PLS.

Application of fractional-order derivative in the quantitative estimation of soil organic matter content through visible and near-infrared spectroscopy

The spectral preprocessing method has become an integral component of soil analysis through visible and near-infrared (Vis-NIR) spectroscopy. Various spectral pretreatment techniques are applied to improve model accuracy by removing undesired side effects, such as spectral noise, baseline shift, and light scattering, and by accentuating spectral features. Conventional integer-order derivatives (i.e., first and second derivatives), which represent specific cases of fractional-order derivatives (FODs), may neglect some detailed spectral information related to target variables. The objective of this study was to compare the performance of FOD in the estimation of soil organic matter (SOM) with that of conventional first and second derivatives. A total of 258 soil samples (180 for calibration and 78 for validation) were collected from Jianghan Plain, Central China. The reflectance spectra and SOM concentrations of the samples were obtained in the laboratory. Two regression techniques, namely, partial least squares (PLS) and PLS–support vector machine (PLS–SVM), and eight FOD transformation processes for spectral data were combined and compared. Results indicated that as the derivative order increased, the details of the FOD spectra changed and the spectral resolution of reflectance curves improved; the intensity of the spectral signals, however, weakened. The correlation between SOM and FOD spectra was enhanced at some specific wavelengths (e.g., the absolute value of the best one-dimensional correlation coefficient of the 1.25-order derivative spectra could reach 0.65 but that of the original reflectance spectra was only 0.47). In most cases, the PLS–SVM models performed better in SOM estimation than the PLS models. The PLS–SVM model with the 1.25-order derivative spectra exhibited the best model performance and provided the validation R2 and ratio of performance to an interquartile range of 0.79 and 3.03, respectively. FOD offers greater advantages in balancing the spectral resolution and the magnitude of spectral strength than traditional integer-order derivatives. Furthermore, the FOD algorithm has strong application potential in soil Vis-NIR spectroscopy and other types of spectroscopy.

Discrimination between abiotic and biotic drought stress in tomatoes using hyperspectral imaging

Crop plants are subjected to various biotic and abiotic stresses. Both root-knot nematodes (biotic stress) and water deficiency (abiotic stress) lead to similar drought symptoms in the plant canopy. In this work, hyperspectral imaging was used for early detection of nematode infestation and water deficiency (drought) stress in tomato plants. Hyperspectral data in the range from 400 to 2500 nm of plants subjected to different watering regimes and nematode infestation levels were analysed by partial least squares – discriminant analysis (PLS-DA) and partial least squares – support vector machine (PLS-SVM) classification. PLS-SVM classification achieved up to 100% accuracy differentiating between well-watered and water-deficient plants, and between 90 and 100% when identifying nematode-infested plants. Grouping the data according to the time of imaging increased the accuracy of classification. Shortwave infrared spectral regions associated with the OH and CH stretches were most relevant for the identification of nematode infested plants and severity of infestation. This study demonstrates the capability of hyperspectral imaging to identify and discriminate between biotic and abiotic plant stresses.

Prediction of Moisture Content for Congou Black Tea Withering Leaves Using Image Features and Nonlinear Method

Withering is the first step in the processing of congou black tea. With respect to the deficiency of traditional water content detection methods, a machine vision based NDT (Non Destructive Testing) method was established to detect the moisture content of withered leaves. First, according to the time sequences using computer visual system collected visible light images of tea leaf surfaces, and color and texture characteristics are extracted through the spatial changes of colors. Then quantitative prediction models for moisture content detection of withered tea leaves was established through linear PLS (Partial Least Squares) and non-linear SVM (Support Vector Machine). The results showed correlation coefficients higher than 0.8 between the water contents and green component mean value (G), lightness component mean value (L*) and uniformity (U), which means that the extracted characteristics have great potential to predict the water contents. The performance parameters as correlation coefficient of prediction set (Rp), root-mean-square error of prediction (RMSEP), and relative standard deviation (RPD) of the SVM prediction model are 0.9314, 0.0411 and 1.8004, respectively. The non-linear modeling method can better describe the quantitative analytical relations between the image and water content. With superior generalization and robustness, the method would provide a new train of thought and theoretical basis for the online water content monitoring technology of automated production of black tea.

Prediction of Soil Organic Matter by VIS–NIR Spectroscopy Using Normalized Soil Moisture Index as a Proxy of Soil Moisture

Soil organic matter (SOM) is an important parameter of soil fertility, and visible and near-infrared (VIS–NIR) spectroscopy combined with multivariate modeling techniques have provided new possibilities to estimate SOM. However, the spectral signal is strongly influenced by soil moisture (SM) in the field. Interest in using spectral classification to predict soils in the moist conditions to minimize the influence of SM is growing. The objective of this study was to investigate the transferability of two approaches, SM–based cluster method with known SM (classifying the VIS–NIR spectra into different SM clusters to develop models separately), the normalized soil moisture index (NSMI)–based cluster method with unknown SM (utilizing NSMI to indicate the SM and establish models separately), to predict SOM directly in moist soil spectra. One hundred and twenty one soil samples were collected from Central China, and eight SM levels were obtained for each sample through rewetting experiments. Their reflectance spectra and SOM concentrations were measured in the laboratory. Partial least square-support vector machine (PLS-SVM) was employed to construct SOM prediction models. Specifically, prediction models were developed for NSMI–based clusters with unknown SM data. The models were assessed through three statistics in the processes of calibration and validation: the coefficient of determination (R2), root mean square error (RMSE) and the ratio of the performance to deviation (RPD). Results showed that the variable SM led to reduced VIS–NIR reflectance nonlinearly across the entire spectral range. NSMI was an effective spectral index to indicate the SM. Classifying the VIS–NIR spectra into different SM clusters in known SM states could improve the performance of PLS-SVM models to acceptable prediction accuracies (R2cv = 0.69–0.77, RPD = 1.79–2.08). The estimation of SOM, when using the NSMI–based cluster method with unknown SM (RPD = 1.95–2.04), was similar to the use of the SM–based cluster method with known SM (RPD = 1.79–2.08). The predictive results (RPD = 1.87–2.06) demonstrated that the NSMI-–based cluster method has potential for application outside the laboratory for SOM prediction without knowing the SM explicitly, and this method is also easy to carry out and only requires spectral information.

Absorption Features in Soil Spectra Assessment.

From a wide range of techniques appropriate to relate spectra measurements with soil properties, partial least squares (PLS) regression and support vector machines (SVM) are most commonly used. This is due to their predictive power and the availability of software tools. Both represent exclusively statistically based approaches and, as such, benefit from multiple responses of soil material in the spectrum. However, physical-based approaches that focus only on a single spectral feature, such as simple linear regression using selected continuum-removed spectra values as a predictor variable, often provide accurate estimates. Furthermore, if this approach extends to multiple cases by taking into account three basic absorption feature parameters (area, width, and depth) of all occurring features as predictors and subjecting them to best subset selection, one can achieve even higher prediction accuracy compared with PLS regression. Here, we attempt to further extend this approach by adding two additional absorption feature parameters (left and right side area), as they can be important diagnostic markers, too. As a result, we achieved higher prediction accuracy compared with PLS regression and SVM for exchangeable soil pH, slightly higher or comparable for dithionite-citrate and ammonium oxalate extractable Fe and Mn forms, but slightly worse for oxidizable carbon content. Therefore, we suggest incorporating the multiple linear regression approach based on absorption feature parameters into existing working practices.

A Multistep Chaotic Model for Municipal Solid Waste Generation Prediction.

In this study, a univariate local chaotic model is proposed to make one-step and multistep forecasts for daily municipal solid waste (MSW) generation in Seattle, Washington. For MSW generation prediction with long history data, this forecasting model was created based on a nonlinear dynamic method called phase-space reconstruction. Compared with other nonlinear predictive models, such as artificial neural network (ANN) and partial least square-support vector machine (PLS-SVM), and a commonly used linear seasonal autoregressive integrated moving average (sARIMA) model, this method has demonstrated better prediction accuracy from 1-step ahead prediction to 14-step ahead prediction assessed by both mean absolute percentage error (MAPE) and root mean square error (RMSE). Max error, MAPE, and RMSE show that chaotic models were more reliable than the other three models. As chaotic models do not involve random walk, their performance does not vary while ANN and PLS-SVM make different forecasts in each trial. Moreover, this chaotic model was less time consuming than ANN and PLS-SVM models.

Simulated annealing based hybrid forecast for improving daily municipal solid waste generation prediction.

A simulated annealing (SA) based variable weighted forecast model is proposed to combine and weigh local chaotic model, artificial neural network (ANN), and partial least square support vector machine (PLS-SVM) to build a more accurate forecast model. The hybrid model was built and multistep ahead prediction ability was tested based on daily MSW generation data from Seattle, Washington, the United States. The hybrid forecast model was proved to produce more accurate and reliable results and to degrade less in longer predictions than three individual models. The average one-week step ahead prediction has been raised from 11.21% (chaotic model), 12.93% (ANN), and 12.94% (PLS-SVM) to 9.38%. Five-week average has been raised from 13.02% (chaotic model), 15.69% (ANN), and 15.92% (PLS-SVM) to 11.27%.

Prediction of CO concentrations based on a hybrid Partial Least Square and Support Vector Machine model

Due to the health impacts caused by exposures to air pollutants in urban areas, monitoring and forecasting of air quality parameters have become popular as an important topic in atmospheric and environmental research today. The knowledge on the dynamics and complexity of air pollutants behavior has made artificial intelligence models as a useful tool for a more accurate pollutant concentration prediction. This paper focuses on an innovative method of daily air pollution prediction using combination of Support Vector Machine (SVM) as predictor and Partial Least Square (PLS) as a data selection tool based on the measured values of CO concentrations.The CO concentrations of Rey monitoring station in the south of Tehran, from Jan. 2007 to Feb. 2011, have been used to test the effectiveness of this method. The hourly CO concentrations have been predicted using the SVM and the hybrid PLS–SVM models. Similarly, daily CO concentrations have been predicted based on the aforementioned four years measured data. Results demonstrated that both models have good prediction ability; however the hybrid PLS–SVM has better accuracy. In the analysis presented in this paper, statistic estimators including relative mean errors, root mean squared errors and the mean absolute relative error have been employed to compare performances of the models. It has been concluded that the errors decrease after size reduction and coefficients of determination increase from 56 to 81% for SVM model to 65–85% for hybrid PLS–SVM model respectively. Also it was found that the hybrid PLS–SVM model required lower computational time than SVM model as expected, hence supporting the more accurate and faster prediction ability of hybrid PLS–SVM model.

Simultaneous characterization of multiple properties of solid and liquid phases in crystallization processes using NIR

Near infrared spectroscopy (NIR) is now probably the most popular process analytical technology (PAT) for pharmaceutical and some other industries. However, unlike mid-IR, NIR is known to have difficulties in monitoring crystallization or precipitation processes because the existence of solids could cause distortion of the spectra. This phenomenon, seen as unfavorable previously, is however an indication that NIR spectra contain rich information about both solids and liquids, giving the possibility of using the same instrument for multiple property characterization. In this study, transflectance NIR calibration data was obtained using solutions and slurries of varied solution concentration, particle size, solid concentration and temperature. The data was used to build calibration models for prediction of the multiple properties of both phases. Predictive models were developed for this challenging application using an approach that combines genetic algorithm (GA) and support vector machine (SVM). GA is used for wavelength selection and SVM for mode building. The new GA–SVM approach is shown to outperform other methods including GA–PLS (partial least squares) and traditional SVM. NIR is thus successfully applied to monitoring seeded and unseeded cooling crystallization processes of l-glutamic acid.