Analytica Chimica Acta 689 (2011) 190–197

Contents lists available at ScienceDirect

Analytica Chimica Acta

journal homepage: www.elsevier.com/locate/aca

Biodiesel classification by base stock type (vegetable oil) using near infrared
spectroscopy data
Roman M. Balabin a,∗ , Ravilya Z. Safieva b
a
Department of Chemistry and Applied Biosciences, ETH Zurich, 8093 Zurich, Switzerland
b
Gubkin Russian State University of Oil and Gas, 119991 Moscow, Russia

a r t i c l e i n f o a b s t r a c t

Article history: The use of biofuels, such as bioethanol or biodiesel, has rapidly increased in the last few years. Near
Received 7 October 2010 infrared (near-IR, NIR, or NIRS) spectroscopy (>4000 cm−1 ) has previously been reported as a cheap and
Received in revised form 17 January 2011 fast alternative for biodiesel quality control when compared with infrared, Raman, or nuclear magnetic
Accepted 19 January 2011
resonance (NMR) methods; in addition, NIR can easily be done in real time (on-line). In this proof-of-
Available online 26 January 2011
principle paper, we attempt to find a correlation between the near infrared spectrum of a biodiesel
sample and its base stock. This correlation is used to classify fuel samples into 10 groups according to
Keywords:
their origin (vegetable oil): sunflower, coconut, palm, soy/soya, cottonseed, castor, Jatropha, etc. Prin-
Petroleum (fossil) fuel
Vegetable (plant) oil
cipal component analysis (PCA) is used for outlier detection and dimensionality reduction of the NIR
Biofuel (biodiesel, bioethanol, spectral data. Four different multivariate data analysis techniques are used to solve the classification
ethanol–gasoline fuel) problem, including regularized discriminant analysis (RDA), partial least squares method/projection on
Vibrational spectroscopy (infrared, latent structures (PLS-DA), K-nearest neighbors (KNN) technique, and support vector machines (SVMs).
near-infrared, and Raman) Classifying biodiesel by feedstock (base stock) type can be successfully solved with modern machine
Artificial neural networks learning techniques and NIR spectroscopy data. KNN and SVM methods were found to be highly effective
Support vector machines for biodiesel classification by feedstock oil type. A classification error (E) of less than 5% can be reached
using an SVM-based approach. If computational time is an important consideration, the KNN technique
(E = 6.2%) can be recommended for practical (industrial) implementation. Comparison with gasoline and
motor oil data shows the relative simplicity of this methodology for biodiesel classification.
© 2011 Elsevier B.V. All rights reserved.

1. Introduction NIR spectroscopic studies have been conducted throughout the

The applications of spectroscopy in various fields of mod-
ern science are increasing every year [1–4]. The most popular
spectroscopic techniques include infrared (IR), near infrared
(NIR/near-IR), and Raman spectroscopies [1–6], ultraviolet–visible
(UV–vis) absorption spectroscopy (spectrophotometry) [1,4], flu-
orescence spectroscopy (fluorometry/spectrofluorometry) [7–10],
and nuclear magnetic resonance spectroscopy (1 H and 13 C NMR)
[11]. Of these, near infrared spectroscopy has several advantages
over other analytical tools because it is non-invasive, requires
minimal sample preparation, and can yield a response in real
time [2,3], which is an important advantage for process analytical
chemistry (PAC) [12]. NIR spectroscopy is based on the absorption
of electromagnetic radiation in the region from 780 to 2500 nm
(12,820–4000 cm−1 ) [3].
∗ Corresponding author. Tel.: +41 44 632 4783.
E-mail addresses: balabin@org.chem.ethz.ch, balabinrm@yandex.ru
(R.M. Balabin).
past few decades. Applications of NIR spectroscopic data can be
found in medical and biomedical studies and imaging [13–16],
nanoscience and nanotechnology (nanotech) [17–20], Earth sci-
ence (geoscience) [21], physical chemistry and chemical physics
[22–25], astronomy [26,27], analytical chemistry [2,3,12,28–30],
food science [3], forestry, and wood science [31,32], and the phar-
maceutical and petroleum industries [22–24,28–30,33,34] (see Ref.
[3] for additional references). When associated with multivariate
data analysis (MDA) [35–40], vibrational spectroscopic techniques
(IR, NIR, and Raman) [1,35–37] have proven to be powerful tools in
the analysis of a variety of fuel samples, including gasoline, diesel,
alcohol fuel (ethanol–gasoline mixtures [41,42]), and kerosene (jet
fuel) [43]. Vibrational spectroscopic techniques are faster than the
usual (ASTM) techniques in the analysis of petroleum refining and
petrochemical products [22,23,44,45]; in addition, these methods
demonstrate good accuracy and precision, are non-destructive, and
can be used in remote quality control [3,12].
Analysis of near-IR spectra usually involves a combination of
multiple samples, each of which has a large number of correlated
features [2,3,12,33]. As such, a variety of data mining algorithms
have been introduced to reduce the complexity accompanying such

0003-2670/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.aca.2011.01.041
 R.M. Balabin, R.Z. Safieva / Analytica Chimica Acta 689 (2011) 190–197
large amounts of data and aim to identify meaningful patterns in
NIR spectra [45–47]. Both regression and classification tasks were
successfully completed using IR/NIR spectral data [48–50].
Increasing energy costs and environmental concerns have
emphasized the need to produce sustainable renewable fuels
and chemicals [51–55]. Ideally, these alternatives should be eco-
nomically competitive, technically achievable, environmentally
acceptable and widely available [51]. The use of renewable ener-
gies, such as biofuels [53], biomass [55], wave, hydro, wind and
solar energies, needs to become widespread in order to decrease
the dependence on fossil fuels [51].
The demand for ethanol and biodiesel, called alternative fuels
or biofuels, has increased over the last few years due to environ-
mental, economical, political, and social issues [44,54]. Biodiesel
is composed of fatty acid mono-alkyl esters, which are obtained
through the base-catalyzed transesterification of vegetable oils or
animal fats with a short chain alcohol, such as methanol (MeOH,
CH3 OH) or ethanol (EtOH, C2 H5 OH) [44]. This biofuel is the major
substitute for petroleum-derived diesel. Because its physical prop-
erties are very similar to diesel, the use of pure or blended biodiesel
does not require any modification of the diesel engine or to the
existing fuel distribution and storage infrastructure [44]. Increased
biodiesel use leads to the reduction of greenhouse gases, particulate
matter, and sulphur emissions. In addition, waste frying oils (WFO)
can be used as a raw material in biodiesel production, thus reducing
and reusing industrial and household waste [51]. Recently, blends
of biodiesel with mineral diesel have become commercially avail-
able world-wide [44,56].
Depending on geographic limitations and oil prices, biodiesel
may be produced from a variety of feedstocks, and different
technologies can be applied for biodiesel production [51–58]. Con-
sequently, the final product can have different properties, so quality
control for biodiesel is very important. The EN 14214 mandates
twenty-five (25) parameters that must be analyzed to certify
biodiesel quality [51].
Near infrared spectroscopy has already been applied to quality
analysis of biodiesel fuel. In 2007, Correia and co-workers [57] pub-
lished their analysis of the effect of commonly used pre-processing
techniques that were applied prior to partial least squares (PLS) and
principal components regression (PCR). They compared the qual-
ity of the developed calibration models to relate the near infrared
spectral characteristics of a biodiesel sample and its methanol and
water content. From the 50 samples studied, 38 were used for cali-
bration and 12 for testing. Leave-one-out cross-validation (LOOCV)
was applied [57].
Chemometric treatment of NIR spectra was assessed by Galtier
et al. [59] for the quantification of fatty acids and triacylglycerols
in virgin olive oil samples. Their classification into five registered
designations of origin (RDOs) of French virgin olive oils, including
“Aix-en-Provence”, “Haute-Provence”, “Nice”, “Nyons” and “Vallee
des Baux”, was successfully conducted using PLS classifiers. It
was concluded [59] that chemometric treatments of NIR spectra
allowed one to obtain similar results than those obtained by time-
consuming analytical techniques, such as GC and HPLC, and these
methods constitute a fast and robust means for authentication of
the French virgin olive oils.
The goals of the study conducted by Yang et al. [60] in 2005 were
to use Fourier transform mid-infrared (FTIR), near-infrared (FT-
NIR) and Raman (FT-Raman) spectroscopy to discriminate among
10 different edible oils and fats and to compare the performance
of these spectroscopic methods. The spectral features of edible oils
and fats were studied, and the characteristic vibrations of C C dou-
ble bond were identified and used for discriminant analysis (DA)
[60].
In this paper, we attempt to find a correlation between the
near infrared spectrum of a biodiesel sample and its feedstock.
191
This correlation is used to correctly classify samples into 10
groups according to their origin (type of vegetable oil) [61]. Prin-
cipal component analysis (PCA) is used for outlier detection and
dimensionality reduction of NIR spectral data [28–30,45–50]. Four
different multivariate data analysis techniques were used to solve
the problem, including regularized discriminant analysis (RDA),
partial least squares method/projection on latent structures (PLS)
[also known as partial least squares-discriminant analysis (PLS-
DA)], K-nearest neighbors (KNN) technique, and support vector
machines (SVMs) [61]. Notably, near infrared spectroscopy in com-
bination with MDA techniques has not been previously applied to
classify biodiesel fuel by its original feedstock.
2. Experimental
2.1. Materials
2.1.1. Industrial samples
Sixty-five samples of rapeseed biodiesel fuel were supplied
by three companies: Yugrostexport (Russia), Biosam (Russia), and
Sokhna Biodiesel Co. (Egypt).
2.1.2. Vegetable oils and used frying oil
Nine types of vegetable oils were used for biodiesel preparation.
Refined and crude sunflower oils, coconut oil, and palm oil were
obtained from Kochmeister (Russia). Soy (soya) bean oil, cottonseed
oil, and castor oil were obtained from MasloPromBaza (Russia). Jat-
ropha (Jatrophacurcas) oil was obtained from Purandhar Agro &
Biofuels (India). Linseed (flax or common flax) oil was obtained
from local commercial sources. The kitchen of a local restaurant
provided the used frying oil. All oil samples were used without
further purification.
2.1.3. Other chemicals
ACS spectroscopic grade methanol (≥99.9%; Sigma–Aldrich),
analytical grade sodium chloride (Sigma–Aldrich), analytical grade
phosphoric acid (Fluka), and reagent grade potassium hydroxide
(≥90%; Sigma–Aldrich) were used throughout the study without
further purification.
2.2. Methods
2.2.1. Biodiesel preparation
The reactor was initially charged with an amount of oil that
achieved the desired methanol/vegetable oil molar ratio, which
ranged from 6:1 to 9:1, and the reactor was then placed in a
constant-temperature bath with its associated equipment and
heated to a temperature of 25–45 ◦ C. The catalyst consisted of
0.4–2.0% (w/w) of potassium hydroxide (KOH) by weight of veg-
etable oil and was completely dissolved in the methanol under
stirring; the solution was then added to the reactor. The timing of
the reaction began as soon as the potassium hydroxide/methanol
solution was added and continued for 3–8 h. The impeller speed
was 600 rpm. Following the reaction, the mixture was transferred
to a separatory funnel, and glycerol was allowed to separate
by gravity for 2–3 h. After the glycerol layer was removed, the
excess methanol in the methyl ester phase was removed by rotary
evaporation at 70 ◦ C. The methyl ester was washed twice with a
phosphoric acid water solution (5%, v/v) and brine until a clear
phase (methyl ester) was obtained.
Outlier detection was done according to Ref. [71]. The final,
“outlier-free” biodiesel sample set included 403 samples from
10 feedstocks: 65 industrial samples from rapeseed oil and 338
laboratory-prepared samples from different oils. The laboratory
samples included sunflower oil (44 samples), coconut oil (40 sam-
ples), palm oil (40 samples), soy oil (40 samples), cottonseed oil (30
192 R.M. Balabin, R.Z. Safieva / Analytica Chimica Acta 689 (2011) 190–197
samples), castor oil (40 samples), Jatropha oil (29 samples), linseed
oil (35 samples), and used frying oil (40 samples). Each sample was
obtained with a separate synthesis.
2.2.2. Near infrared (NIR) spectroscopy
The near infrared (NIR) spectra were acquired with a MPA Multi
Purpose FT-NIR Analyzer (Bruker). The spectra were acquired at
room temperature (20–23 ◦ C). The NIR spectrometer was calibrated
with benzene (C6 H6 ) and cyclohexane (c-C6 H12 ) at least twice per
day to minimize the influence of variable laboratory conditions.
The spectral range between 9000 and 4500 cm−1 (1110–2500 nm)
was scanned with a resolution of 8 cm−1 . Sixty-four scans were
averaged for each sample spectrum, and a background spectrum
(64 scans) was measured every 45 min. A photometric accuracy of
∼0.05% was obtained [2,28–30]. A cylindrical glass cell was used
throughout the study. Approximately 1 mL of biodiesel sample was
needed for each NIR measurement. NIR spectrum collection was
repeated five times with cell rotation inside the spectrometer to
minimize interferences from the cell or glass defects. The mea-
surement of one sample took less than 5 min. The averaged- and
background-corrected spectra were used for subsequent data pre-
processing [28–30].
2.2.3. Near infrared (NIR) spectral pre-processing
Data pre-processing was done according to Refs. [51,57].
Briefly, prior to calibration model building, various widely used
pre-processing techniques were applied to the data. Eight data
pre-treatment methods were tested:
• MC: Mean Centering;
• MSC-MC: Mean Scattering Correction followed by Mean Center-
ing;
• SNV-MC: Standard Normal Variate scaling plus Mean Centering;
• SGD1-MC: First-order Savitzky–Golay derivative followed by
Mean Centering;
• SGD2-MC: Second-order Savitzky–Golay derivative followed by
Mean Centering;
• MC-OSC: Mean Centering followed by the Orthogonal Signal Cor-
rection method;
• SGD1-MC-OSC: SGD1-MC followed by the Orthogonal Signal Cor-
rection method;
• SGD2-MC-OSC: SGD2-MC followed by the Orthogonal Signal Cor-
rection method.
The SGD2-MC-OSC method produced the best results and was
used for data pre-treatment, unless otherwise specified.
2.2.4. Classifiers: RDA, PLS-DA, KNN, and SVM
2.2.4.1. Regularized discriminant analysis (RDA). Regularized dis-
criminant analysis (RDA) can be called a “compromise” between
linear discriminant analysis (LDA) and quadratic discriminant anal-
ysis (QDA). RDA introduces a double type of bias in the estimation
of the covariance matrices of each class with the aim of stabilizing
the prediction in the case of a deficient estimation of their elements.
This is a very frequent problem encountered with real data because
the estimations of the matrices are very sensitive to the presence
of different data. For a detailed analysis of RDA behavior one can
consult Ref. [62].
2.2.4.2. Partial least squares regression/classification (PLS-DA). Par-
tial least squares regression (PLS-DA) is an extension of the multiple
linear regression model and principal component analysis (PCA)
method. The PLS-DA method has found wide-spread use for multi-
variate analysis of spectral data. The main goal of PLS-DA usage is
calibration model building, but this technique can also be applied
for classification purposes. For PLS-DA classification, the class (one
of M) of each sample is coded with a (binary) vector of length M
with zeros and 1 one. In our case, M is equal to 10. The M predicted
scores by the ordinary PLS-DA method are used to predict the class
of each sample. If the j-score (normalized) is greater than or equal
to parameter ı, the predicted class of the sample was j. Extra infor-
mation about the idea of PLS-DA classification can be found in Ref.
[63].
2.2.4.3. K-nearest neighbor (KNN). The K-nearest neighbor (KNN)
method was first introduced by Fix and Hodges [64]. It has an
extremely simple, intuitive theoretical background. In the KNN
method, a distance (Euclidean or other, see below) is assigned
between all points in a data set. The data points, which are K clos-
est neighbors (K being the number of neighbors), are then found by
analyzing (sorting) the distance matrix. The K-closest data points
are then analyzed to determine which class label is the most com-
mon among the set (“voted” KNN). The most common class label
is then assigned to the data point being analyzed. Other variants
of KNN (“weighted” KNN) are also possible. In these methods, dif-
ferent neighbors have different “weight”, which is the ability to
influence the class of a sample, according to their distances from
the sample. Some remarks about the possible types of functions
for weighting are given below (Table 1). For a KNN classifier, it is
necessary to have a training set that is not too small and a good
discriminating distance. KNN performs well in multi-class, simul-
taneous problem solving. There exists an optimal choice for the
value of the parameter K, which describes the best performance of
the classifier. For a detailed analysis, one can consult Ref. [64].
2.2.4.4. Support vector machines (SVMs). Support vector machines
were developed by Vapnik [65]. Support vector machines (SVMs)
are based on some “beautifully simple ideas” [66] and provide a
clear demonstration of what learning from examples is all about.
SVMs are also showing strong performances in practical appli-
cations [65–70]. In very simple terms, an SVM corresponds to a
linear method in a very high-dimensional feature space that is
nonlinearly related to the input space. Even though we think of
it as a linear algorithm in a high-dimensional feature space, SVM
does not involve any computations in that high-dimensional space
in practice. Through the use of kernels (Table 2), all necessary
computations are performed directly in the input space. Support
vector machines map input vectors to a higher dimensional space
where a maximal separating hyperplane is constructed. Two par-
allel hyperplanes are constructed on each side of the hyperplane
that separates the data. The hyperplane is the hyperplane that
maximizes the distance between the two parallel hyperplanes. An
assumption is made that the larger the margin or distance between
these parallel hyperplanes, the better the generalization error of the
classifier will be. Details about SVM classifiers can be found in Refs.
[65,66].
2.2.5. Data analysis
2.2.5.1. Software and computing. The initial spectra were digitised
using the special software complex created by one of the authors
(B.R.M.) [28]. After digitisation, each spectrum was represented
as a 1 × 563 vector; the length of the vector was defined by the
spectrometer resolution. The software package MATLAB 2008b
(Mathworks, Inc., Natick, MA) along with the Statistics Toolbox and
the Neural Network Toolbox were extensively used in designing
and executing multivariate procedures. Standard programs were
modified and extended by B.R.M. Self-written software was used
for spectral pre-processing [28–30].
2.2.5.2. Model efficiency (calibration error) estimation. To character-
ize prediction ability (efficiency) of created classification model the
 R.M. Balabin, R.Z. Safieva / Analytica Chimica Acta 689 (2011) 190–197 193

Table 1
Parameters of the K-nearest neighbor (KNN) classification model.

Weighting Formulaa Parameters Number of neighbors (K)

b
None – – K = 1–29
Weighting 1 w(xi ) = exp(−D(xi , p)) – K = 1–29
−q
Weighting 2 w(xi ) = D(xi , p) q = 1–5 K = 1–25
Weighting 3 w(xi ) = 1
2 ε = 10−3 –102 K = 1–20
D(xi ,p) +ε
−1
Weighting 4 w(xi ) = (D(xi , p) + 1) – K = 1–29

1, D(xi , p) ≤ d0
Weighting 5 w(xi ) = d0 = 10−4 –102 K = 1–20
0, D(xi , p) > d0

 of distances D(xi , p) between vectors xi and p were used: Euclidean distance (Pythagorean metric):
a
Two types
 L

D(xi , p) = 
2
(xi,l − pl )
l=1

and Manhattan distance (taxicab geometry):


L

D(xi , p) = |xi,l − pl |.
l=1
b
“Voting” KNN.

error of cross-validation (E) was used:
Nwrong N0 − Nright
E= = (1)
N0 N0
where Nwrong and Nright refer to the number of wrongly and
rightly classified samples; N0 is the total number of samples in
validation set. The value of E is a measure of model sensitivity
(SENS = 100% − E), i.e., the ability to classify in a correct manner
the objects belonging to the class (minimization of false negatives).
However, there is another fundamental parameter, specificity
(SPEC), that expresses the ability of a given class model to correctly
reject the objects of other classes (minimization of false positives)
[69,70]. In our case (Tables 3–6) almost no difference is observed if
any error definition is used (e.g., mean of SENS and SPEC) for mul-
tivariate methods comparison. So in this work we follow the same
error function (E) as in Ref. [61] and references therein. Note that
one can calculate any other error type from the data in Tables 3–6,
if he or she prefers other functional for methods comparison.
Fivefold (5-fold) cross-validation was used to evaluate model’s
efficiency. It was checked that validation set consists of all biodiesel
classes. The results of cross-validation check were found to be
almost identical to the test ones on a fully independent sample set.
So, cross-validation (CV) error will be used as a default measure of
model accuracy, if other is not specified.
2.2.5.3. Efficiency estimation with external test set. To check the
validity of the CV procedure the tests with external test sets were
done. The full sample set was randomly divided into three subsets:
calibration (283), cross-validation test set/test set for parameter
optimization (50), and external test sets (70). It was checked that
all classes are presented in each set. The test set has 6–10 samples
of every oil type. The procedure has been repeated five times and
average values are reported.
2.2.5.4. Model optimisation. To truly compare different multivari-
ate classification methods, the efficiency of the best possible model
should be found. Because of the dependence of the calibration
model efficiency on its parameters, the following parameters were
varied and optimised using cross-validation error (RMSE) [28–30]:
• RDA: one parameter (˛ = 0–1);
• PLS-DA: number of latent variables (LV = 1–35) and threshold
parameter (ı = 0.40–0.95);
• KNN: different types of K-nearest neighbor (KNN) method (voted
and weighted by different weighting functions) were used to get
the best result (Table 1), K was optimised on each case;
• SVM: different types of kernel functions (and corresponding
parameters) were checked for optimization of SVM classification
model (Table 2), cost parameter C was optimised in each case.
3. Results and discussion
3.1. Biodiesel samples set and its classes
All biodiesel samples in the set can be divided into two broad
categories: biodiesel fuel produced from vegetable oils (363 sam-
ples) and biodiesel fuel produced from used frying oil (40 samples).
A real variety of biodiesel feedstocks (vegetable oils) is presented
in the set: nine types of vegetable oils that are very different in
chemical composition and properties [51–60]. One can expect that
this sample set makes the final model general enough to be used in
real-world applications. Of course, additional samples are needed
if samples from other feedstock are needed to be classified. But the
general trend is believed to stay the same as presented below.
At least 29 samples per class: 40.3 ± 9.9 (±) on average—are
presented in the set. This value can be called acceptable [38–40]. It
makes it possible to create a robust model, since approximately the
same number of samples (±25%) is observed in all cases [40,65]. It
also means that the classification error can be calculated in a direct
way, see Eq. (1)—no correction for sample number difference and/or
making copies of samples is needed [38–40,65].
One would also expect that 403 samples is enough to build
an accurate model at “sample set limit”—compare with “basis set
limit” or “complete basis set” in quantum chemistry (QC) [35,67]. It

Table 2
Parameters of the support vector machine (SVM) classification model.

Kernel Formula Kernel parameters Cost parameter (C)

Linear K(xi , xj ) = xi xj + 1 – C = 10−3 –102

Polynomial K(xi , xj ) = (xi xj + 1)p p = 2–8 C = 10−3 –102
 
||xi −xj ||2
Gaussian K(xi , xj ) = exp −  = 10−2 –102 C = 10−4 –102
2
194 R.M. Balabin, R.Z. Safieva / Analytica Chimica Acta 689 (2011) 190–197

Table 3
Biodiesel classification by regularized discriminant analysis (RDA): confusion matrix. The rows indicate the true sample class and that the columns refer to the observed
class.

Biodiesel type Total number Sunflower Coconut Palm Rapeseed Soy Cottonseed Castor Jatropha Linseed Used frying
of samples oil

Sunflower 44 33 1 1 3 2 1 1 1 1
Coconut 40 1 33 3 1 1 1
Palm 40 1 2 34 2 1
Rapeseed 65 3 60 1 1
Soy 40 1 1 37 1
Cottonseed 30 1 1 28
Castor 40 3 37
Jatropha 29 3 25 1
Linseed 35 3 3 1 2 1 25
Used frying oil 40 1 1 38

Table 4
Biodiesel classification by partial least squares (PLS) classification model: confusion matrix. The rows indicate the true sample class and that the columns refer to the observed
class.

Biodiesel type Total number Sunflower Coconut Palm Rapeseed Soy Cottonseed Castor Jatropha Linseed Used frying
of samples oil

Sunflower 44 33 2 2 1 2 1 1 1 1
Coconut 40 1 35 3 1
Palm 40 3 34 1 1 1
Rapeseed 65 1 63 1
Soy 40 2 37 1
Cottonseed 30 1 1 28
Castor 40 2 38
Jatropha 29 2 27
Linseed 35 3 3 1 1 1 1 25
Used frying oil 40 40

Table 5
Biodiesel classification by K-nearest neighbor (KNN) classification model: confusion matrix. The rows indicate the true sample class and that the columns refer to the observed
class.

Biodiesel type Total number Sunflower Coconut Palm Rapeseed Soy Cottonseed Castor Jatropha Linseed Usedfrying
of samples oil

Sunflower 44 38 2 2 2
Coconut 40 40
Palm 40 3 33 2 1 1
Rapeseed 65 64 1
Soy 40 1 1 35 2 1
Cottonseed 30 2 28
Castor 40 40
Jatropha 29 29
Linseed 35 1 1 1 1 31
Used frying oil 40 40

Table 6
Biodiesel classification by Support vector machine (SVM) classification model: confusion matrix. The rows indicate the true sample class and that the columns refer to the
observed class.

Biodiesel type Total number Sunflower Coconut Palm Rapeseed Soy Cottonseed Castor Jatropha Linseed Usedfrying
of samples oil

Sunflower 44 40 1 1 2
Coconut 40 39 1
Palm 40 1 2 35 1 1
Rapeseed 65 64 1
Soy 40 38 2
Cottonseed 30 30
Castor 40 40
Jatropha 29 29
Linseed 35 1 1 1 1 1 30
Used frying oil 40 40

is well-known that the accuracy of a multivariate model increases
with the increasing size of the sample set and saturates at a limit
value at very large sample sets. The same effect is observed in ab ini-
tio quantum chemistry for basis set—set of Gaussian atom-cantered
functions used to create the molecular orbitals and to approximate
electron density distribution in molecular systems.
3.2. Biodiesel classification by regularized discriminant analysis
(RDA)
Table 3 shows the results of regularized discriminant analysis
(RDA) of biodiesel fuels. One can see that the model is not accu-
rate enough for practical use. A classification error of 13.2% was
 R.M. Balabin, R.Z. Safieva / Analytica Chimica Acta 689 (2011) 190–197 195

Table 7
Classification of biodiesel samples by different multivariate methods: regularized discriminant analysis (RDA), partial least squares classification (PLS), K-nearest neighbor
(KNN), and support vector machines (SVMs).

RDA PLS KNN SVM

Nright /N0 E Nright /N0 E Nright /N0 E Nright /N0 E

Sunflower 33/44 25% 33/44 25% 38/44 14% 40/44 9%

Coconut 33/40 18% 35/40 13% 40/40 0% 39/40 3%
Palm 34/40 15% 34/40 15% 33/40 18% 35/40 13%
Rapeseed 60/65 8% 63/65 3% 64/65 2% 64/65 2%
Soy 37/40 8% 37/40 8% 35/40 13% 38/40 5%
Cottonseed 28/30 7% 28/30 7% 28/30 7% 30/30 0%
Castor 37/40 8% 38/40 5% 40/40 0% 40/40 0%
Jatropha 25/29 14% 27/29 7% 29/29 0% 29/29 0%
Linseed 25/35 29% 25/35 29% 31/35 11% 30/35 14%
Used frying oil 38/40 5% 40/40 0% 40/40 0% 40/40 0%

Total 350/403 13.2% 360/403 10.7% 378/403 6.2% 385/403 4.5%

E is the error of 5-fold cross-validation, see Eq. (1); Nright refers to the number of wrongly and rightly classified samples; N0 is the total number of samples in validation set.
Methods: RDA: regularized discriminant analysis; PLS: partial least squares regression; KNN: K-nearest neighbor; and SVMs: support vector machines.

reached based on 350 correctly classified biodiesel samples out of
403.
In the cases of biodiesel from sunflower oil and linseed oil, an
error above 25% was observed (Table 3). The best results were
observed for the samples from the used frying oil; E = 5% for 38 cor-
rectly classified biodiesel samples out of 40. The confusion matrix
(Table 3) shows that there is a significant chemical difference in
the feedstock (used frying oil vs. vegetable oil), which makes clas-
sification among these two species rather simple, and the problem
can easily be solved by a simple RDA model. The same cannot be
said about classification among vegetable oils; a more sophisticated
classification method is needed in this case.
3.3. Biodiesel classification by the partial least squares (PLS-DA)
classification model
The confusion matrix for the partial least squares or projection to
latent structures (PLS-DA) classification model (Table 4) shows that
the PLS-DA method results are close to those for the RDA method.
A classification error of 10.7% is obtained (360 correctly classified
biodiesel samples out of 403). This is better than the RDA results
(see above), but these results are still inadequate to state that the
model is highly accurate.
Interestingly, two of the most problematic types of samples for
classification yield the same results as in the RDA method: sun-
flower and linseed oils. In both cases, E is above 25%. An accuracy
of 100% (40/40) is reached in the case of the used frying oil clas-
sification. Thus, the problem of a vegetable oil vs. used frying oil
distinction is completely solved at the PLS-DA level.
Though PLS-DA classification has been shown to be a more effec-
tive classification method than standard RDA, more sophisticated
classification methods are needed to build highly accurate models
for biodiesel classification by feedstock type.
3.4. Biodiesel classification by the K-nearest neighbors (KNN)
classification model
The K-nearest neighbors (KNN) classification model is a sim-
ple but highly effective classification algorithm [64]; this fact is
confirmed by this study using a biodiesel sample set (Table 5). A
classification error of only 6.2% was reached by an application of the
KNN model. The largest inaccuracy (E = 18%) is observed for palm
oil. In the cases of coconut, castor, Jatropha, and used frying oils,
100% effectiveness in classification using this model was reached.
This final model based on K-nearest neighbors classification
can be recommended for practical implementation as 378 sam-
ples were successfully assigned to their correct classes. This level
of accuracy is significantly better than the accuracy achieved using
the PDA or PLS-DA methods (see above).
3.5. Biodiesel classification by support vector machines (SVMs)
Table 6 summarises the results of the SVM approach to biodiesel
classification. The confusion matrix (Table 6) clearly shows the
superiority of a support vector approach over all of the classification
methods described above [29,68]. The extent of the classification
error decreases from 6.2% to 4.5% for 385 correctly classified sam-
ples at a much higher computational cost (see below).
The largest inaccuracy (E = 14%) is observed for linseed oil. In
the cases of cottonseed, castor, Jatropha, and used frying oils, 100%
effectiveness in classification using this model was reached, which
is nearly the same result provided by the KNN method. Importantly,
not more than two samples were labelled incorrectly by a SVM-
based approach.
Support vector machines can be regarded as the most effective
classification model for biodiesel fuels. Very accurate (E < 5%) and
robust classification models can be built using a SVM approach.
3.6. General remarks. Comparison with gasoline and motor oil
data
Table 7 summarises the data presented above for a general dis-
cussion. Table 7 clearly shows the non-equivalency of the classes in
term of classification accuracy. Three types of vegetable oil have an
average classification error above 15%: sunflower, palm, and linseed
oils. Thus, biodiesel samples from these three groups are difficult to
label correctly. Tables 4–6 show that this inaccuracy is mostly based
on the confusion between these three classes. The general accuracy
of a classification model seems to be greatly dependent on its ability
to correctly distinguish sunflower-, palm-, and linseed-based fuel
samples. Only the KNN and SVM methods are rather successful in
this task.
Three other types of biodiesel samples may be described as
“easy-to-distinguish”. These are: rapeseed, castor, and used frying
oils. In these cases, the average error is below 4%. The difference
in chemical composition [51–60] and, as a consequence, the differ-
ence in near infrared spectral features [58–60] may be responsible
for this attribute.
Comparison of the effectiveness of RDA, PLS-DA, KNN, and SVM
for different fuel and oil samples (gasoline, motor oil, and biodiesel)
is presented in Fig. 1 [29,68]. Motor oils of 10 popular brands (Cas-
trol, Mobil, Esso, TNK-BP, Shell, Total, Lukoil, BP, Mannol, Liqui
moly; SAE types: 0W-20; 5W-20; 0W-30; 5W-30; 10W-30; 0W-40;
5W-40; 10W-40; 15W-40; 5W-50; 10W-50) were used for classifi-
196 R.M. Balabin, R.Z. Safieva / Analytica Chimica Acta 689 (2011) 190–197
Fig. 1. Comparison of the efficiency of classification models (classification error)
for three sample sets: gasoline [29], motor oil [70], and biodiesel. RDA: regularized
discriminant analysis; PLS-DA: partial least squares method/projection on latent
structures; KNN: K-nearest neighbors; SVM: support vector machines.
cation [29]. Motor oil samples (210–225 items) were classified into
3–4 classes according to base stock (also origin) and viscosity. Three
sets of gasoline near infrared (NIR) spectra (450, 415, and 345 spec-
tra) were used for classification of gasoline samples into 3–6 classes,
according to their source (refinery or process) and type [68]. These
data can also be regarded as gasoline origin. The 14,000–8000 cm−1
NIR spectral region was chosen for classification. All of the classifi-
cations are based on near infrared spectroscopic data.
One can see that the general trend that relates classification
model order with respect to accuracy is the same in all three
cases: RDA > PLS-DA > KNN > SVM. In all of the cases, RDA and PLS-
DA methods yield approximately the same error of classification,
which is relatively large.
The main difference between products of petroleum refining
and petrochemicals (gasoline and motor oil) and renewable energy
sources (biodiesel) is the accuracy ratio between PLS-DA/KNN
and KNN/SVM methods. In the case of gasoline and motor oil, a
SVM-based approach shows significant superiority over RDA- and
PLS-DA-based ones. The difference is not that pronounced in the
case of biodiesel classification (Fig. 1). The reason for this could be
in the relative simplicity of biodiesel as a chemical system when
compared to gasoline and especially motor oil. Notably, the dif-
ference in near-IR spectral ranges [29,68] and number of classes
[29,68] may be partially responsible for the differences that are
observed.
The choice between KNN- and SVM-based approaches is also
dependent on the computational resources available. While the
KNN algorithm is easy to create and use, SVM-based methods are
rather computationally demanding in terms of CPU time, and the
computer resources needed, in the model construction and opti-
mization phase. In case of online use, once the models are made,
the differences are small. For off-line quality control of feedstocks,
these differences are not as important as for on-line one.
Note that for on-line applications, it is only necessary to assess
the class of a given sample using a previously calculated model.
Therefore, the time spent to calculate the model is not impor-
tant, since this work has already been done previously (calibration
model building). In this application phase, there is only the need
to assign the sample to the proper class, and the time necessary to
perform this operation is independent of the time needed to cal-
culate the models and to select the optimal one. So, it is only the
building of the SVM model that takes a long time.
3.7. Tests with independent (external) test set
All multivariate results should be tested using fully indepen-
dent (external) data sets, which were not used in the process of
calibration model building. Only such data can be regarded as true
estimation of model real-world performance.
Since the selection of many parameters (e.g., kernel function
and cost parameter in SVM) is based on the minimum CV error,
the classification methods that seem to give the best performances
are those, where the number of free parameters is higher. It is
well known that an increased number of possible combinations of
parameters leads to increased possibilities of chance in classifica-
tion models [40].
The independent set of 70 biodiesel samples was used (see
Section 2.2.5). The results (E) of RDA, PLS-DA-DA, KNN, and SVM
classifications are 14.6 ± 0.6% (±), 10.6 ± 1.3%, 6.9 ± 1.2%, and
4.6 ± 0.6%, respectively. These results are very close to the results
of 5-fold cross-validation procedure, reported earlier. So, in the
case of biodiesel classification CV and external results are very
close. A small error increase in case of test sample set usage can be
explained by, both, the underestimation of error by CV procedure
and the decrease of the calibration sample set by ∼12%.
Of course, the same situation (ECV ≈ Etest ) cannot be expected in
all other cases when near-infrared spectroscopy is used to classify
fuel samples, and independent tests are always necessary.
Note that the CV values reported (Tables 3–7) can be directly
compared to the motor oil and gasoline data [29,68] (see Section
3.6). The main point here is to prove that the CV error is close to the
test one (see above).
4. Conclusions
The following conclusions can be drawn from this study:
(1) The task of biodiesel classification by feedstock (base stock)
type can be successfully solved by modern machine learning
techniques based on near infrared (NIR) spectroscopic data.
(2) K-nearest neighbors (KNN) and support vector machine (SVM)
methods were found to be highly effective for biodiesel classi-
fication by feedstock oil type (Fig. 1).
(3) A classification error below 5% can be reached by a SVM-based
approach. The error per class lies in the range of 0–14%.
(4) If computational time is an important consideration, the KNN
technique can be recommended (E = 6.2%) for practical (indus-
trial) implementation.
(5) Comparison with gasoline and motor oil data demonstrates the
relative simplicity of this biodiesel classification task.
Further work has to be done to proof that the proposed tech-
nique is generally applicable for different biodiesel feedstocks
(from different suppliers or from different origins of the same type
of vegetable oil). The results presented herein can be applied for the
rapid and accurate analysis of other biofuels, products of petroleum
and petrochemicals. The use of vibrational spectroscopy such as
near infrared spectroscopy in other fields of analytical chemistry,
including pharmaceutical (drug) quality control, food quality con-
trol, and analysis of tablets, can be enhanced by the application
of modern methods of multivariate data analysis [69–73]. We
believe that both classification and regression tasks for industrially
important systems can be solved using modern machine learning
(computer based) techniques.
Acknowledgements
B.R.M. is grateful to the ITERA International Group of companies
for a scholarship. Bruker Optics Inc. (Moscow, Russia) is acknowl-
edged for the use of their equipment in the preliminary studies.
References
[1] J.M. Hollas, Modern Spectroscopy, 4th ed., Blackwell, Wiley, 2003.
 R.M. Balabin, R.Z. Safieva / Analytica Chimica Acta 689 (2011) 190–197
[2] J. Workman, A. Springsteen, Applied Spectroscopy: A Compact Reference for
Practitioners, Academic Press, 1998.
[3] B. Osborne, T. Fearn, Near Infrared Spectroscopy in Food Analysis, Wiley, New
York, 1986.
[4] F.J. Duarte, Tunable Laser Applications, CRC, New York, 2009.
[5] F.J. Duarte, Tunable Lasers Handbook, Academic Press, 1995.
[6] W. Demtröder, Laser Spectroscopy: Basic Principles, 4th ed., Springer, Berlin,
2008.
[7] J.R. Lakowicz, Principles of Fluorescence Spectroscopy, Kluwer Aca-
demic/Plenum Publishers, 1999.
[8] A. Sharma, S.G. Schulman, Introduction to Fluorescence Spectroscopy, Wiley,
1999.
[9] J. Logan, K. Edwards, N. Saunders, Real-Time PCR: Current Technology and
Applications, Caister Academic Press, 2009.
[10] J. Eisinger, J. Flores, Anal. Biochem. 94 (1979) 15–21.
[11] J. Keeler, Understanding NMR Spectroscopy, John Wiley & Sons, 2005.
[12] J. Workman, M. Koch, B. Lavine, R. Chrisman, Anal. Chem. 81 (2009) 4623–4643.
[13] A. Burke, X. Ding, R. Singh, R.A. Kraft, N. Levi-Polyachenko, M.N. Rylander, C.
Szot, C. Buchanan, J. Whitney, J. Fisher, H.C. Hatcher, R. D’Agostino, N.D. Kock,
P.M. Ajayan, D.L. Carroll, S. Akman, F.M. Torti, S.V. Torti, Proc. Natl. Acad. Sci. U.
S. A. 106 (2009) 12897–12902.
[14] E.R. Trivedi, A.S. Harney, M.B. Olive, I. Podgorski, K. Moin, B.F. Sloane, A.G.M.
Barrett, T.J. Meade, B.M. Hoffman, Proc. Natl. Acad. Sci. U. S. A. 107 (2010)
1284–1288.
[15] X. Michalet, F.F. Pinaud, L.A. Bentolila, J.M. Tsay, S. Doose, J.J. Li, G. Sundaresan,
A.M. Wu, S.S. Gambhir, S. Weiss, Science 307 (2005) 538–544.
[16] L.R. Hirsch, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 13549–13554.
[17] M.J. Therien, Nature 458 (2009) 716–717.
[18] T.F. Krauss, R.M. De La Rue, S. Brand, Nature 383 (1996) 699–702.
[19] Y. Zou, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 22135–22138.
[20] R. Adato, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 19227–19232.
[21] H. Keppler, L.S. Dubrovinsky, O. Narygina, I. Kantor, Science 322 (2008)
1529–1532.
[22] R.M. Balabin, R.Z. Safieva, J. Near Infrared Spectrosc. 15 (2007) 343–347.
[23] R.M. Balabin, R.Z. Syunyaev, J. Colloid Interface Sci. 318 (2008) 167–174.
[24] R.Z. Syunyaev, R.M. Balabin, I.S. Akhatov, J.O. Safieva, Energy Fuels 23 (2009)
1230–1238.
[25] O.C. Mullins, Anal. Chem. 62 (1990) 508–514.
[26] M.R. Swain, G. Vasisht, G. Tinetti, Nature 452 (2008) 329–331.
[27] S. Byrne, Science 325 (2009) 1674–1676.
[28] R.M. Balabin, R.Z. Safieva, E.I. Lomakina, Chemometr. Intell. Lab. Syst. 88 (2007)
183–188.
[29] R.M. Balabin, R.Z. Safieva, Fuel 87 (2008) 1096–1101.
[30] R.M. Balabin, R.Z. Safieva, E.I. Lomakina, Chemometr. Intell. Lab. Syst. 93 (2008)
58–64.
[31] L.R. Schimleck, R. Evans, A.C. Matheson, J. Wood Sci. 48 (2002) 132–137.
[32] P.D. Jones, L.R. Schimleck, G.F. Peter, R.F. Daniels, A. Clark, Wood Sci. Technol.
40 (2006) 709–720.
[33] E.W. Ciurczak, J.K. Drennen, Pharmaceutical and Medicinal Applications of
Near-infrared Spectroscopy, 1st ed., CRC Press, 2002.
[34] J. Moros, N. Galipienso, R. Vilches, S. Garrigues, M. de la Guardia, Anal. Chem.
80 (2008) 7257–7264.
[35] P. Taddei, S. Affatato, C. Fagnano, B. Bordini, A. Tinti, A. Toni, J. Mol. Struct. 613
(2002) 121–129.
197
[36] M. Andersson, K.-G. Knuuttil, Vib. Spectrosc. 29 (2002) 133–138.
[37] L.M. Harwood, C.J. Moody, Experimental Organic Chemistry: Principles and
Practice, Wiley-Blackwell, 1989.
[38] T. Næs, T. Isaksson, T. Fearn, T. Davies, A User-Friendly Guide to Mul-
tivariate Calibration and Classification, NIR Publications, Chichester, UK,
2002.
[39] R.M. Balabin, E.I. Lomakina, J. Chem. Phys. 131 (2009) 074104.
[40] B.F.J. Manly, Multivariate Statistical Methods: A Primer, 3rd ed., Chapman and
Hall/CRC, 2004.
[41] R.M. Balabin, R.Z. Syunyaev, S.A. Karpov, Fuel 86 (2007) 323–327.
[42] R.M. Balabin, R.Z. Syunyaev, S.A. Karpov, Energy Fuels 21 (2007) 2460–2465.
[43] S.B. Kim, C. Temiyasathit, K. Bensalah, A. Tuncel, J. Cadeddu, W. Kabbani, A.V.
Mathker, H. Liu, Expert Syst. Appl. 37 (2010) 3863–3871.
[44] M.R. Monteiroa, A.R.P. Ambrozin, M.S. Santos, E.F. Boffo, E.R. Pereira-Filho, L.M.
Lião, A.G. Ferreira, Talanta 78 (2009) 660–664.
[45] D.D. Lee, H.S. Seung, Nature 401 (1999) 788–790.
[46] J.B. Tenenbaum, V. de Silva, J.C. Langford, Science 290 (2000) 2319–2321.
[47] G.E. Hinton, R.R. Salakhutdinov, Science 313 (2006) 504–507.
[48] D.C. Malins, N.L. Polissar, S.J. Gunselman, Proc. Natl. Acad. Sci. U. S. A. 94 (1997)
3611–3615.
[49] K. Sakurai, Y. Goto, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 15346–15351.
[50] A.E. Cohen, W.E. Moerner, Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 12622–12627.
[51] P. Baptista, P. Felizardo, J.C. Menezes, M.J.N. Correia, Talanta 77 (2008)
144–151.
[52] E.J. Steen, Nature 463 (2010) 559–562.
[53] A.J. Ragauskas, Science 311 (2006) 484–489.
[54] A.E. Farrell, R.J. Plevin, B.T. Turner, A.D. Jones, M. O’Hare, D.M. Kammen, Science
311 (2006) 506–508.
[55] S. Crossley, J. Faria, M. Shen, D.E. Resasco, Science 327 (2009) 68–72.
[56] G. Knothe, J. Am. Oil Chem. Soc. 78 (2001) 1025–1028.
[57] P. Felizardo, P. Baptista, J.C. Menezes, M.J.N. Correia, Anal. Chim. Acta 595 (2007)
107–113.
[58] F.C.C. Oliveira, C.R.R. Brandao, H.F. Ramalho, L.A.F. Costa, P.A.Z. Suarez, J.C.
Rubim, Anal. Chim. Acta 587 (2007) 194–199.
[59] O. Galtier, N. Dupuya, Y. Le Dreau, D. Ollivier, C. Pinatel, J. Kister, J. Artaud, Anal.
Chim. Acta 595 (2007) 136–144.
[60] H. Yang, J. Irudayaraj, M.M. Paradkar, Food Chem. 93 (2005) 25–32.
[61] R.M. Balabin, R.Z. Safieva, E.I. Lomakina, Anal. Chim. Acta 671 (2010) 27–35.
[62] J.H. Friedman, J. Am. Stat. Assoc. 84 (1989) 16–1755.
[63] P. Geladi, B.R. Kowalski, Anal. Chim. Acta 185 (1986) 1–17.
[64] E. Fix, J.L. Hodges, Int. Stat. Rev. 57 (1989) 238.
[65] V.N. Vapnik, The Nature of Statistical Learning Theory, Springer-Verlag, New
York, 1995.
[66] S.R. Amendolia, G. Cossu, M.L. Ganadu, B. Golosio, G.L. Masala, G.M. Mura,
Chemometr. Intell. Lab. Syst. 69 (2003) 13–20.
[67] R.M. Balabin, J. Chem. Phys. 129 (2008) 164101.
[68] R.M. Balabin, R.Z. Safieva, Fuel 87 (2008) 2745–2752.
[69] L. Pigani, G. Foca, A. Ulrici, K. Ionescu, V. Martina, F. Terzi, M. Vignali, C. Zanardi,
R. Seeber, Anal. Chim. Acta 643 (2009) 67–73.
[70] L. Pigani, G. Foca, K. Ionescu, V. Martina, A. Ulrici, F. Terzi, M. Vignali, C. Zanardi,
R. Seeber, Anal. Chim. Acta 614 (2008) 213–222.
[71] T. Lillhonga, P. Geladi, Anal. Chim. Acta 544 (2005) 177–183.
[72] R.M. Balabin, R.Z. Safieva, E.I. Lomakina, Microchem. J., in press.
[73] R.M. Balabin, E.I. Lomakina, R.Z. Safieva, Fuel, in press.