Multiblock Methodology Research Articles

Multiblock Analysis Applied to TD-NMR of Butters and Related Products

This work presents a novel and rapid approach to predict fat content in butter products based on nuclear magnetic resonance longitudinal (T1) relaxation measurements and multi-block chemometric methods. The potential of using simultaneously liquid (T1L) and solid phase (T1S) signals of fifty samples of margarine, butter and concentrated fat by Sequential and Orthogonalized Partial Least Squares (SO-PLS) and Sequential and Orthogonalized Selective Covariance Selection (SO-CovSel) methods was investigated. The two signals (T1L and T1S) were also used separately with PLS and CovSel regressions. The models were compared in term of prediction errors (RMSEP) and repeatability error (σrep). The results obtained from liquid phase (RMSEP ≈ 1.33% and σrep≈ 0.73%) are better than those obtained with solid phase (RMSEP ≈ 5.27% and σrep≈ 0.69%). Multiblock methodologies present better performance (RMSEP ≈ 1.00% and σrep≈ 0.47%) and illustrate their power in the quantitative analysis of butter products. Moreover, SO-Covsel results allow for proposing a measurement protocol based on a limited number of NMR acquisitions, which opens a new way to quantify fat content in butter products with reduced analysis times.

NIR and MIR spectral data fusion for rapid detection of Lonicera japonica and Artemisia annua by liquid extraction process

Lonicera japonica and Artemisia annua are two widely used plants that are typically combined together as botanical dietary supplements or herbal medicines. In this study, the liquid extraction process of various mixtures of these two plants was investigated with near-infrared (NIR) and mid-infrared (MIR) spectroscopy for rapid detection of important effective compounds. Eight main effective compounds of Lonicera japonica and Artemisia annua mixtures were modeled and selected as quality attributes: neochlorogenic acid (5-CQA), cryptochlorogenic acid (4-CQA), chlorogenic acid (3-CQA), caffeic acid (CA), isochlorogenic acid B (ICAB), isocholorogenic acid A (ICAA), isocholorogenic acid C (ICAC), and secoxyloganin (SL). First, partial least squares (PLS) was adopted to build quantitative models by combining spectra from MIR and NIR instruments. Two multiblock PLS methods were applied using NIR and MIR data matrix in which concatenated PLS (C-PLS) and sequential and orthogonalized PLS (SO-PLS) were utilized to optimize model performance. The results showed that models based on single NIR and MIR spectroscopy had close capacity for detection of the eight indexes. However, performance of the fusion methods C-PLS and SO-PLS both achieved better results than the results of single NIR or MIR PLS models, with SO-PLS being superior to C-PLS. This indicates that the multiblock methodology can be used to optimize rapid detection of indexes from complex industrial processes of plants using NIR and MIR spectroscopy.

Establishing the optimal blocks' order in SO‐PLS: Stepwise SO‐PLS and alternative formulations

AbstractMethods capable of handling multiple blocks of data have increased in popularity both in industry and academia, namely, in the fields of chemistry, food and beverages, pharmacology, and biology. Among the multiblock methodologies, sequential orthogonalized partial least squares (SO‐PLS) has been attracting great interest, given its interpretation capabilities (eg, the possibility to estimate the relative additional contributions of each block to predict the response and the degree of mutual overlap, ie, the blocks' communalities) associated to desirable modeling features, such as the independence from the relative scaling of the different data blocks and the flexibility to handle blocks with different dimensionalities and pseudo‐ranks. Given the sequential nature of SO‐PLS, it is critically dependent on the order of the blocks used. When a priori knowledge exists about the natural ordering of the blocks (eg, data arising from sequential operations in a production process), this specification is straightforward. However, in the absence of such knowledge or in cases where no order should be preferred a priori (as happens in the case study of this article), SO‐PLS faces the problem of having to find the most adequate one through the exhaustive search of all permutations. This problem is particularly relevant when the number of blocks is larger than 3, getting exponentially worse as this number increases. Situations where the number of natural data blocks is significant are already quite frequent (eg in multistage batch operations) and will tend to occur more often in the future, as Manufacturing 4.0 takes its course and data sources from the entire value chain become more abundant. Therefore, more efficient and systematic approaches are required to retain the benefits of SO‐PLS while coping with the increasing demands for data processing and analysis in Big Data applications. In this article, we introduce Stepwise SO‐PLS as an efficient algorithm for selecting the blocks ordering when performing SO‐PLS, with capabilities of block exclusion (a feature not shared by other current multiblock approaches). A robust statistical comparison framework based on Monte Carlo cross‐validation is employed to compare the proposed stepwise SO‐PLS formulation with the current systematic approach for selecting the block order and its new variant allowing for blocks selection, in their prediction capabilities. A real case study of the prediction of Madeira wine age will be used for establishing the comparison.

Mid-infrared and near-infrared spectroscopy for rapid detection of Gardeniae Fructus by a liquid-liquid extraction process

Gardeniae Fructus is widely used in the pharmaceutical industry, and many studies have confirmed its medical and economic value. In this study, samples collected from different liquid-liquid extraction batches of Gardeniae Fructus were detected by mid-infrared (MIR) and near-infrared (NIR) spectroscopy. Seven analytes, neochlorogenic acid (5-CQA), cryptochlorogenic acid (4-CQA), chlorogenic acid (3-CQA), geniposidic acid (GEA), deacetyl-asperulosidic acid methyl ester (DAAME), genipin-gentiobioside (GGB), and gardenoside (GA), were chosen as quality property indexes of Gardeniae Fructus. The two kinds of spectra were each used to build models by single partial least squares (PLS). Additionally, both spectral data were combined and modeled by multiblock PLS. For single spectroscopy modeling results, NIR had a better prediction for high-concentration analytes (3-CQA, DAAME, GGB, and GA) whereas MIR performed better for low-concentration analytes (5-CQA, 4-CQA, and GEA). The multiblock methodology was found to be better compared to single spectroscopy models for all seven analytes. Specifically, the coefficients of determination (R2) of the NIR, MIR, and multiblock PLS calibration models of all seven components were higher than 0.95. Relative standard errors of prediction (RSEP) were all less than 7%, except for models of GGB, which were 10.36%, 13.24%, and 8.15% for the NIR-PLS, MIR-PLS, and multiblock models, respectively. These results indicate that MIR and NIR spectrographic techniques could provide a new choice for quality control in industrial production of Gardeniae Fructus.

Advanced predictive methods for wine age prediction: Part II – A comparison study of multiblock regression approaches

In this article, we extend the scope of the first paper of the sequel, which was dedicated to the analysis of advanced single-block regression methods (Rendall et al., 2016) [1], to the class of multiblock regression approaches. The datasets contemplated for developing the multiblock approaches are the same as in the former publication: volatile, polyphenols, organic acids composition and the UV–Vis spectra. The context is still the prediction of the ageing time of one of finest Portuguese fortified wines, the Madeira Wine, but now the data collected from the different analytical sources is explored simultaneously, in a more structured and informative way, through multiblock methodologies. The goal of this paper is to provide a critical assessment of a rich variety of multiblock regression methods, namely: Concatenated PLS, Multiblock PLS (MBPLS), Hierarchical PLS (HPLS), Network-Induced Supervised Learning (NI-SL) and Sequential Orthogonalised Partial Least Squares (SO-PLS). A comparison of block scaling methods was also undertaken for the Concatenated PLS algorithm, and new block scaling methods were proposed that led to better prediction performances. This study explores and reveals the potential advantages of applying multiblock methods for fusing datasets from different sources, from both the predictive and interpretability perspectives.

Complex multiblock bottle-brush architectures by RAFT polymerization.

The combination of the reversible addition fragmentation chain transfer (RAFT) polymerization R-group grafting from approach and RAFT one-pot acrylamide multiblock methodology is used to synthesise complex bottle-brush architectures. This is demonstrated for block copolymer grafted side chains and the insertion of ungrafted blocks into the backbone to yield multi-segmented bottle-brushes.

A sensory approach for the monitoring of accelerated red wine aging processes using multi-block methods

Two young red wines from D.O. Navarra (Spain) were separately subjected to accelerated aging in tanks with oak chips added and micro-oxygenation. Wine samples were collected during each process and were analyzed by a sensory panel according to 17 descriptors. A two-fold monitoring approach, aimed at both checking panel performance and tracking the impact of chips/micro-oxygenation treatments on sensory properties, was proposed for process control and optimisation. In a first stage this approach focused on investigating panel performance by suitably combining uni- and multivariate methods. Significant product effects were found for 7 and 6 attributes in the first and second aging series, respectively. Although panel performance was acceptable in both cases, moderate differences between assessors were detected, stressing the need for adopting a consensus approach to properly handle them. Thus, the STATIS method was subsequently applied to find an optimal compromise solution for the significant attributes describing each process. The resulting compromise matrix provided the consensus sensory profiles of samples at different aging stages. The assignment of individual weights was consistent with previous findings on assessors’ agreement. The first principal component of the compromise matrix, which accounted for 86.5% and 90.9% of the variance in each trial data, revealed a sample trend directly related to the time-course evolution of each process, and contrasted the attributes fruit and herbaceous with the other significant descriptors (related to the wood–wine interaction). This is the first time that a multi-block methodology has been used for monitoring sensory changes during accelerated aging processes.

Incorporating interactions in multi-block sequential and orthogonalised partial least squares regression

This paper is about how to incorporate interaction effects in multi-block methodologies. The method proposed is inspired by polynomial regression modelling in the case with only a few independent variables but extends/generalises the idea to situations where the blocks are potentially very large with respect to the number of variables. The method follows a so-called type I sums of squares strategy where the linear effects (main effects) are incorporated sequentially and before the interactions. The sequential and orthogonalised partial least squares (SO-PLS) technique is used as a basis for the proposal. The SO-PLS method is based on sequential estimation of each new block by the PLS regression method after orthogonalisation with respect to blocks already fitted. The new method preserves the invariance already established for SO-PLS and can be used for blocks with different dimensionality. The method is tested on one real data set with two independent blocks with different complexity and on a simulated data set with a large number of variables in each block. Copyright © 2011 John Wiley & Sons, Ltd.

Evaluation of multiblock NIR/MIR PLS predictive models to detect adulteration of diesel/biodiesel blends by vegetal oil

This work investigates the potential of using simultaneously near infrared (NIR) and mid infrared (MIR) spectroscopy for the quantification of vegetable oil in diesel/biodiesel blends. The two spectral ranges were used separately with PLS regressions. To combine the two pieces of information, first a concatenated matrix was used and then H-PLS and S-PLS models were constructed. The models were compared in term of prediction errors (RMSEP). The results obtained in NIR spectroscopy were better than the ones obtained with MIR spectroscopy (considering the influence of the pretreatment and of the selected variables range). The multiblock methodology seems to be of great interest in quantitative analysis with the simultaneous use of information from the MIR and NIR spectroscopy.

Identifying outlying assessors in sensory profiling using fuzzy clustering and multi-block methodology

In this paper we discuss methods for detecting outlying assessors in descriptive sensory analysis. A new method is proposed which is based on fuzzy clustering with the use of the noise cluster method. The technique ends up with a plot which can be used to provide information about which assessors that are different from the rest and also to suggest reasons why the assessors are different. The method is based on data compression by the use of either Tucker-1 or Tucker-2 for multi-block data. The technique is used on a dataset based on an oxidation experiment of cheese. There are 9 trained and 3 untrained assessors in the panel.

Multiblock PLS as an approach to compare and combine NIR and MIR spectra in calibrations of soybean flour

The present work is aimed at investigating the potential benefits of simultaneously combining near-infrared (NIR) and mid-infrared (MIR) spectral regions for use in calibration development of soybean flour quality properties (crude protein and moisture). NIR and MIR spectra were analysed separately using single partial least squares (PLS), and then both spectral data sets were utilized together in the modelling by applying two multiblock methodologies based on PLS regression: Multiblock PLS (MB-PLS) and Serial PLS (S-PLS). Utilizing the concept of net analyte signal (NAS), models constructed from NIR or MIR data were compared in terms of analytical figures of merit (sensitivity (SEN), selectivity (SEL) and limit of detection (LOD)). When utilized alone, the MIR spectra gave models with considerably inferior prediction power and analytical figures of merit than NIR-based models. The multiblock methodology revealed to be very useful, since it helped to determine if there was distinctive information in each spectral data set and evaluate the relative importance of each data set. The results pointed out the existence of additional information in the MIR spectra not present in the NIR spectra. Although several works have already been reported comparing NIR and MIR spectroscopic techniques using different multivariate regression techniques, at the best of our knowledge none of them applied the approach of multiblock PLS. Therefore, the present work intends to explore this direction, using the NIR and MIR spectra as predictor blocks to model flour's properties in a parallel mode (MB-PLS) or in a serial mode (S-PLS).

Novel multi-block strategy for CAD tools for microfluidics type applications

A novel strategy to implement multi-block schemes in Computer aided design (CAD) tools for a variety of applications specifically dealing with multi-physics transport phenomena is presented here. Multi-block schemes are domain decomposition schemes in which the domain is split into several smaller sub-domains, and independent, synchronous calculations are performed in the sub-domains individually, retaining the global identity of the problem. These schemes are very appropriate for problems with complex geometries and physics, and problems that are computationally intensive. They also allow parallelization on distributed and shared memory parallel architectures. Therefore, for complex multi-physics applications like microfluidics, which is spotlighted here, multi-block strategies can offer many advantages over traditional finite element or finite volume schemes. However, from a CAD tool implementation point of view multi-block schemes are still in infancy. One important reason for this is the different kind of geometry and mesh builders required for multi-block schemes owing to the inherently different philosophy entailed. To circumvent this, in this work we present some novel strategies to utilize the commonly used geometry builders and meshing tool of finite element technique for multi-block solvers, and develop a versatile CAD tool based on domain decomposition technique. In particular a lucid methodology is presented using which rudimentary and widely available geometry builder and meshing tool for finite element calculations can be utilized to implement a powerful multi-block scheme using structured grids and finite-volume discretization, both of which are amenable to ease of implementation and numerical efficiency for the class of problems considered. The proposed multi-block methodology can also be used with finite difference or finite element calculations in individual blocks. It will be shown that this formulation will have the potential to be the bulwark of next generation of micro- and nano-scale fluidics type problems that have applications in areas like proteomics, lab-on-a-chip design, synthetic biomaterials, etc.