degradacion ac. ascorbico

12
Neural network prediction of ascorbic acid degradation in green asparagus during thermal treatments Hong Zheng, Shuangshuang Fang, Heqiang Lou, Yong Chen, Lingling Jiang, Hongfei Lu College of Chemistry and Life Science, Zhejiang Normal University, Jinhua 321004, China a r t i c l e i n f o Keywords: Articial neural network Ascorbic acid Kinetic  Asparagus offcinalis  L. Thermal treatments a b s t r a c t An articial neural network was developed to predict the kinetics of ascorbic acid loss in green asparagus during thermal treatments and the model was trained using a back-propagation algorithm. The results indicate that the optimal ANN models consisted one hidden layer and the optimal number of neurons in the hidden layer was 24, 26, 26 and 18 for bud, upper, middle and butt segments of asparagus, respec- tively. The ANNs could predict the kinetic parameters of ascorbic acid degradation in asparagus with an MSE of 1.3925 and MAE 0.528 3 for bud segment, MSE 2.4618 and MAE 0.6436 for upper segment, MSE 0.8985 and 0.4258 for middle segment and MSE 0.2707 and MAE 0.1883 for butt segment. In addition, the correlation coefcients between experimental  k,  t 1/2  or  D-value and predicted values were greater than 0.99 in all cases. Therefore, ANN offers a simple, quick and convenient means of the kinetic parameters prediction in chemical kinetics.  2010 Elsevier Ltd. All rights reserved. 1. Introduction Asparagus (  Asparagus offcinalis L., family  Liliaceae) is a healthy and nutritious vegetable, containing antioxidants, such as rutin, ascorbic acid, tocopherol, ferulic acid and glutathione ( Shao et al., 1997). Ascorbic acid (AA) is an important component of our nutri- tion and used as additive in many foods because of its antioxidant capacity, however, it is known to be thermolabile and is easily de- stroyed during processing, especially thermal treatments ( Garrote, Silva , & Bertone, 1986). Blanching is one of the many processes that take place during the preparation of raw vegetables before preser- vation processes like canning and freezing. Quantitative represen- tation of kinetic data has also been extensively reported in an attemp t to predic t and optimize ascorbic acid retenti on during blanching and storage. To date, several authors have studied the kinetics of AA degradation in foods during blanching and storage and stated that it follows a rst-order kinetic model ( Al-Zubaidy & Khalil, 2007; Arroqui, Rumsey, Lopez, & Virseda, 2002; Burdurlu, Koca, & Karadeniz, 2006; Huelin, 1953; Johnson, Braddock, & Chen, 1995; Lee & Coates, 1999; Vieira, Teixeira, & Silva, 2000 ). Articial neural network (ANN) is a set of mathe matica l meth- ods, often encompassed with articial intelligence, which in some way attempt to mimic the functioning of the human brain ( Bila et al., 1999). Recently, interest in using ANN as a modeling tool in food technology is increasing. ANN has been successfully used in several food applications like sensory analysis, classications, micro bial predicti ons or therma l contr ol among others ( Afaghi, Ramaswamy, & Prashe r, 2001; Geeraerd, Herremans , Cenen s, & Van Impe, 1998; Guyer & Yang, 2000; Ni & Gunasekaran, 1998; Ruan, Almaer, & Zhang, 1995). Unlik e other modeling techniqu es such as simultaneo us heat and mass transf er, kineti c models, and regression analysis, an ANN can accommodate more than two vari- ables to predict two or more parameters. In this study, we have developed an articial neural network useful to predict kinetics of ascorbic acid degradation in asparagus during blanching at dif- ferent temperatures. Many authors have proposed mathematical models and computer simulations representing ascorbic acid con- centration as a function of moisture content, temperature and time (Haralampu & Karel, 1983; Mishkin, Saguy, & Karel, 1984; Villota & Karel, 1980). However, to our knowledge no previous study has been reported on using ANNs to predict the kinetics of ascorbic acid degradation. The present work was undertaken to: (1) evaluating the kinetics of ascorbic acid degradation during blanching in different parts of asparagus and (2) develop the kinetics of ascorbic acid loss predic- tion models based on articial neural networks. 2. Materials and methods  2.1. Thermal treatments Fresh asparagus (  Asparagus offcinalis  L. var. Grande) was har- vested from a local farm in Jinhua (Zhejiang, PR China) and trans- ported by refrigeration at 8 C for 30 min to the laborat ory. Spears of the same diameter (0.8–1.0 cm) at the base and length (20 cm) 0957-4174/$ - see front matter  2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2010.10.076 Corresponding author. Tel.: +86 0579 8228 2284. E-mail address:  [email protected] (H. Lu). Expert Systems with Applications 38 (2011) 5591–5602 Contents lists available at  ScienceDirect Expert Systems with Applications journal homepage:  www.elsevier.com/locate/eswa

Upload: erick-ten

Post on 04-Jun-2018

223 views

Category:

Documents


0 download

TRANSCRIPT

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 1/12

Neural network prediction of ascorbic acid degradation in green asparagus

during thermal treatments

Hong Zheng, Shuangshuang Fang, Heqiang Lou, Yong Chen, Lingling Jiang, Hongfei Lu ⇑

College of Chemistry and Life Science, Zhejiang Normal University, Jinhua 321004, China

a r t i c l e i n f o

Keywords:

Artificial neural network

Ascorbic acid

Kinetic

 Asparagus offcinalis L.

Thermal treatments

a b s t r a c t

An artificial neural network was developed to predict the kinetics of ascorbic acid loss in green asparagusduring thermal treatments and the model was trained using a back-propagation algorithm. The results

indicate that the optimal ANN models consisted one hidden layer and the optimal number of neuronsin the hidden layer was 24, 26, 26 and 18 for bud, upper, middle and butt segments of asparagus, respec-tively. The ANNs could predict the kinetic parameters of ascorbic acid degradation in asparagus with an

MSE of 1.3925 and MAE 0.5283 for bud segment, MSE 2.4618 and MAE 0.6436 for upper segment, MSE0.8985 and 0.4258 for middle segment and MSE 0.2707 and MAE 0.1883 for butt segment. In addition, the

correlation coefficients between experimental  k,  t 1/2  or  D-value and predicted values were greater than0.99 in all cases. Therefore, ANN offers a simple, quick and convenient means of the kinetic parameters

prediction in chemical kinetics. 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Asparagus ( Asparagus offcinalis L., family  Liliaceae) is a healthyand nutritious vegetable, containing antioxidants, such as rutin,ascorbic acid, tocopherol, ferulic acid and glutathione (Shao et al.,

1997). Ascorbic acid (AA) is an important component of our nutri-tion and used as additive in many foods because of its antioxidantcapacity, however, it is known to be thermolabile and is easily de-stroyed during processing, especially thermal treatments (Garrote,

Silva, & Bertone, 1986). Blanching is one of the many processes thattake place during the preparation of raw vegetables before preser-vation processes like canning and freezing. Quantitative represen-

tation of kinetic data has also been extensively reported in anattempt to predict and optimize ascorbic acid retention duringblanching and storage. To date, several authors have studied thekinetics of AA degradation in foods during blanching and storage

and stated that it follows a first-order kinetic model (Al-Zubaidy& Khalil, 2007; Arroqui, Rumsey, Lopez, & Virseda, 2002; Burdurlu,Koca, & Karadeniz, 2006; Huelin, 1953; Johnson, Braddock, & Chen,1995; Lee & Coates, 1999; Vieira, Teixeira, & Silva, 2000 ).

Artificial neural network (ANN) is a set of mathematical meth-ods, often encompassed with artificial intelligence, which in someway attempt to mimic the functioning of the human brain ( Bilaet al., 1999). Recently, interest in using ANN as a modeling tool

in food technology is increasing. ANN has been successfully usedin several food applications like sensory analysis, classifications,

microbial predictions or thermal control among others (Afaghi,

Ramaswamy, & Prasher, 2001; Geeraerd, Herremans, Cenens, &

Van Impe, 1998; Guyer & Yang, 2000; Ni & Gunasekaran, 1998;Ruan, Almaer, & Zhang, 1995). Unlike other modeling techniquessuch as simultaneous heat and mass transfer, kinetic models, and

regression analysis, an ANN can accommodate more than two vari-ables to predict two or more parameters. In this study, we havedeveloped an artificial neural network useful to predict kineticsof ascorbic acid degradation in asparagus during blanching at dif-

ferent temperatures. Many authors have proposed mathematicalmodels and computer simulations representing ascorbic acid con-centration as a function of moisture content, temperature and time

(Haralampu & Karel, 1983; Mishkin, Saguy, & Karel, 1984; Villota &Karel, 1980). However, to our knowledge no previous study hasbeen reported on using ANNs to predict the kinetics of ascorbicacid degradation.

The present work was undertaken to: (1) evaluating the kineticsof ascorbic acid degradation during blanching in different parts of asparagus and (2) develop the kinetics of ascorbic acid loss predic-tion models based on artificial neural networks.

2. Materials and methods

 2.1. Thermal treatments

Fresh asparagus ( Asparagus offcinalis  L. var. Grande) was har-vested from a local farm in Jinhua (Zhejiang, PR China) and trans-

ported by refrigeration at 8  C for 30 min to the laboratory. Spearsof the same diameter (0.8–1.0 cm) at the base and length (20 cm)

0957-4174/$ - see front matter    2010 Elsevier Ltd. All rights reserved.doi:10.1016/j.eswa.2010.10.076

⇑ Corresponding author. Tel.: +86 0579 8228 2284.

E-mail address:  [email protected] (H. Lu).

Expert Systems with Applications 38 (2011) 5591–5602

Contents lists available at   ScienceDirect

Expert Systems with Applications

j o u r n a l h o m e p a g e :   w w w . e l s e v i e r . c o m / l o c a t e / e s w a

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 2/12

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 3/12

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 4/12

Fig. 4.  Ascorbic acid loss in different segments of asparagus during thermal treatments: (a) bud segment; (b) upper segment; (c) middle segment and (d) butt segment.

 Table 2

Kinetic parameters for the thermal degradation of ascorbic acid in bud segment of asparagus during blanching at different temperatures.

Temperature (C) Variation kinetics   k (min1)   t 1/2  (min)   D-value (min)   R2

60   y ¼  462:40expð0:0057 xÞ   0.0057 121.60 175.44 0.9931

65   y ¼  430:73expð0:0086 xÞ   0.0086 80.60 116.28 0.9324

70   y ¼  432:59expð0:0112 xÞ   0.0112 61.89 89.29 0.9450

75   y ¼  422:01expð0:0139 xÞ   0.0139 49.87 71.94 0.8623

80   y ¼  423:62expð0:0267 xÞ   0.0267 25.96 37.45 0.8952

85   y ¼  432:16expð0:0459 xÞ   0.0459 15.10 21.79 0.944090   y ¼  427:85expð0:0734 xÞ   0.0734 9.44 13.62 0.9223

95   y ¼  433:

54expð0:

1592 xÞ   0.1592 4.35 6.28 0.9464100   y ¼  433:30expð0:3040 xÞ   0.3040 2.28 3.29 0.9283

 Table 3

Kinetic parameters for the thermal degradation of ascorbic acid in upper segment of asparagus during blanching at different temperatures.

Temperature (C) Variation kinetics   k (min1)   t 1/2  (min)   D-value (min)   R2

60   y ¼  437:90expð0:0046 xÞ   0.0046 150.68 217.39 0.932465   y ¼  371:60expð0:0081 xÞ   0.0081 85.57 123.46 0.9694

70   y ¼  362:33expð0:0109 xÞ   0.0109 63.59 91.74 0.9834

75   y ¼  359:81expð0:0163 xÞ   0.0163 42.52 61.35 0.989580   y ¼  350:27expð0:0281 xÞ   0.0281 24.67 35.59 0.9801

85   y ¼  343:77expð0:0468 xÞ   0.0468 14.81 21.37 0.9506

90   y ¼  335:50expð0:0763 xÞ   0.0763 9.08 13.11 0.8952

95   y ¼  335:12expð0:1510 xÞ   0.1510 4.59 6.62 0.8951

100   y ¼  340:35expð0:2750 xÞ   0.2750 2.52 3.64 0.9207

5594   H. Zheng et al. / Expert Systems with Applications 38 (2011) 5591–5602

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 5/12

 Table 4

Kinetic parameters for the thermal degradation of ascorbic acid in middle segment of asparagus during blanching at different temperatures.

Temperature (C) Variation kinetics   k (min1)   t 1/2  (min)   D-value (min)   R2

60   y ¼  285:91expð0:0025 xÞ   0.0025 277.26 400.00 0.929265   y ¼  281:65expð0:0039 xÞ   0.0039 177.73 256.41 0.9310

70   y ¼  274:60expð0:0074 xÞ   0.0074 93.67 135.14 0.8947

75   y ¼  271:79expð0:0102 xÞ   0.0102 67.96 98.04 0.8540

80   y ¼  271:60expð0:0216 xÞ   0.0216 32.09 46.30 0.8888

85   y ¼  265:50expð0:0362 xÞ   0.0362 19.15 27.62 0.831290   y ¼  260:20expð0:0609 xÞ   0.0609 11.38 16.42 0.8774

95   y ¼  270:72expð0:1496 xÞ   0.1496 4.63 6.68 0.8846

100   y ¼  274:91expð0:2579 xÞ   0.2579 2.69 3.88 0.9147

 Table 5

Kinetic parameters for the thermal degradation of ascorbic acid in butt segment of asparagus during blanching at different temperatures.

Temperature (C) Variation kinetics   k (min1)   t 1/2  (min)   D-value (min)   R2

60   y ¼  234:86expð0:0029 xÞ   0.0029 239.02 344.83 0.962765   y ¼  229:87expð0:0042 xÞ   0.0042 165.04 238.10 0.8975

70   y ¼  222:26expð0:0061 xÞ   0.0061 113.63 163.93 0.8106

75   y ¼  218:05expð0:0106 xÞ   0.0106 65.39 94.34 0.806780   y ¼  226:55expð0:0217 xÞ   0.0217 31.94 46.08 0.9293

85   y ¼  222:73expð0:0345 xÞ   0.0345 20.09 28.99 0.8848

90   y ¼  215:93expð0:0546 xÞ   0.0546 12.70 18.32 0.8071

95   y ¼  218:30expð0:1298 xÞ   0.1298 5.34 7.70 0.8403

100   y ¼  223:10expð0:2445 xÞ   0.2445 2.83 4.09 0.8812

Fig. 5.  Arrhenius plots of ascorbic acid degradation in bud segment (a), upper segment (b), middle segment (c) and butt segment (d).

H. Zheng et al. / Expert Systems with Applications 38 (2011) 5591–5602   5595

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 6/12

contents of different segments of asparagus were plotted as a func-tion of blanching time at various temperatures (Fig. 4). The AA con-

tent in asparagus decreased with an increase in blanching time andtemperature. Therefore, there was a difference in AA content due todifferent parts of asparagus, blanching temperature and time.

The fitted exponential curves (R2 = 0.8067  0.9931) showedgood results for dependence of AA concentration during blanching.

Therefore, the loss of AA in asparagus at all temperatures is de-scribed as a first-order reaction. The kinetic parameters of AA deg-radation during thermal treatments at different temperatures areshown in Tables 2–5. The t 1/2  and  D-value decreased with increase

in blanching temperature. However, the reduction in the rate con-stant (k) corresponded to a decrease in temperature.

 3.2. Temperature dependence

Fig. 5 shows Arrhenius plots of AA degradation in different seg-ments of asparagus. Activation energies (E a) and  Q 10  values of AA

loss were calculated in asparagus and given in  Table 6. The E a val-ues for AA degradation were 24.24 kcal mol1 for bud segment,

24.68 kcal mol

1

for upper segment, 28.70 kcal mol

1

for middlesegment and 27.56 kcal mol1 for butt segment. The  Q 10  values at60–100 C ranged from 1.96 to 4.14, from 2.37 to 3.60, from 2.96to 4.23 and from 2.10 to 4.48 in bud, upper, middle and butt seg-ments, respectively.   Table 6  showed that the highest   Q 10  values

were observed at 90–100  C, and the lowest  E a values were locatedin bud segment of asparagus.

 3.3. ANN for kinetic parameters prediction

An artificial neural network based on back propagation wasused to predict kinetic parameters of AA degradation in asparagusduring thermal treatments. In our study, an one-hidden-layer FFBPwas used and the number of neurons in the hidden layer varied

from one to thirty.   Fig. 6   illustrated the MSE, MRE, STDR, MAEand STD A, which were used to compare the performances of 

 Table 6

Activation energies  ðE aÞ  and temperature quotient (Q 10) values for ascorbic acid degradation in different segments of asparagus.

E a  (kcal mol1)   Q 10

60–70 C 70–80  C 80–90  C 90–100  C

Bud segment 24.24 1.96 2.38 2.75 4.14

Upper segment 24.68 2.37 2.58 2.72 3.60

Middle segment 28.70 2.96 2.91 2.82 4.23Butt segment 27.56 2.10 3.56 2.52 4.48

Fig. 6.  Errors in the prediction of kinetic parameters with different number of neurons in the hidden layer for bud (a), upper (b), middle (c) and butt segments of asparagus

during blanching.

5596   H. Zheng et al. / Expert Systems with Applications 38 (2011) 5591–5602

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 7/12

various ANN models. The optimum number of hidden layer neu-rons was 24, 26, 26 and 18 for bud, upper, middle and butt seg-

ments of asparagus, respectively (Fig. 6). Plots of experimentallydetermined   k,   t 1/2   or   D-value versus ANN predicted values areshown in Figs. 7–10. The correlation coefficients were greater than0.99 in all cases. For bud segment,  R2 = 1 for predicted  k, 0.9947 for

t 1/2  and 0.9995 for D-value. For upper segment, R2 = 0.9998, 0.9995

and 1 for   k,   t 1/2  and  D-value, respectively. The correlation coeffi-cients ranged from 0.9992 to 0.9999 and from 0.9996 to 1 formiddle and butt segments, respectively. Errors in the prediction of 

k, t 1/2  and D-value of AA degradation in asparagus with the optimalANN were presented in Table 7.

4. Discussion

Vegetables are a major source of ascorbic acid, a nutrient thatbesides its vitamin action is valuable for its antioxidant effect.

Asparagus is a green vegetable with high antioxidant activityamong the commonly consumed vegetables (Vinson, Hao, Su, &

Zubik, 1998). Our data showed that the content of AA was highest

Fig. 7.  Correlation of experimental and predicted kinetic parameters with testing and training data sets (a,c,e), as well as validation data set (b,d,f) for bud segment of asparagus during thermal treatments using the optimal ANN.

H. Zheng et al. / Expert Systems with Applications 38 (2011) 5591–5602   5597

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 8/12

in bud segment of asparagus (0.91 ± 0.04 mg/100 ml juice),followed by upper segment (0.72 ± 0.06 mg/100 ml juice), middlesegment (0.58 ± 0.06 mg/100 ml juice) and was least in butt seg-

ment (0.48 ± 0.04 mg/100 ml juice). However,  Nindo, Sun, Wang,Tang, and Powers (2003) reported that the middle and basal partscontained more ascorbic acid than the tip portion of asparagus.

Although AA is an important component of our nutrition, it is the

least stable of all vitamins andis easily destroyed duringprocessing,especially thermal treatments, because of its thermolability. Garrote

et al. (1986) observed that the main mechanisms of AA loss duringthe blanching operation are thermal induced degradation or byleaching. Ourresults demonstrated that theAA content in asparagus

decreased depending on the blanching time and temperature(Fig. 4). Moreover, AA degradation in asparagus during blanchinghas been described by a first-order reaction model in our experi-ment. In fact, thefirst-order kinetic model hasbeen applied by many

researchers (Frias & Oliveira, 2001; Frias, Oliveira, Cunha,& Oliveira,1998; Giannakourou & Taoukis, 2003; Johnson et al., 1995; Uddin,

Fig. 8.   Correlation of experimental and predicted kinetic parameters with testing and training data sets (a,c,e), as well as validation data set (b,d,f) for upper segment of 

asparagus during thermal treatments using the optimal ANN.

5598   H. Zheng et al. / Expert Systems with Applications 38 (2011) 5591–5602

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 9/12

Hawlader, Ding, & Mujumdar, 2002) for evaluating AA degrada-tion in biological materials of food system. According to  Table 6,

high activation energies and   Q 10   values in middle and butt seg-ments indicated that AA degradation was more temperaturedependent than the other segments. The lowest   E a   value for AAdegradation was obtained in bud segment of asparagus. The low-

est activation energies for AA degradation are also remarkablesince this reaction is favoured at low temperatures. In addition,the   t 1/2  and  D-value of AA loss in bud segment were lower thanthe other segments at the same blanching temperature (Tables

2–5). Therefore, bud segment is more liable to lose AA than theother segments.

From a nutritional point of view, the extent of AA retention is awidely adopted quality criterion. Therefore, many authors have

proposed mathematical models and computer simulations predict-ing ascorbic acid loss during processing and storage (Haralampu &Karel, 1983; Mishkin et al., 1984; Vieira et al., 2000; Villota & Karel,1980). Recently, artificial neural network (ANN) has generated

increasing acceptance and is an interesting method in several foodprocessing applications like sensory analysis and quality control(Bucinski, Zielinski, & Kozłowska, 2004; Cabrera & Prieto, 2010;Lewis et al., 2008; O’Farrell, Lewis, Flanagan, Lyons, & Jackman,

2005; Panigrahi, Balasubramanian, Gu, Logue, & Marchello, 2006),classifications (Gestal et al., 2004; Nadal, Espinosa, Schuhmacher,

Fig. 9.  Correlation of experimental and predicted kinetic parameters with testing and training data sets (a,c,e), as well as validation data set (b,d,f) for middle segment of 

asparagus during thermal treatments using the optimal ANN.

H. Zheng et al. / Expert Systems with Applications 38 (2011) 5591–5602   5599

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 10/12

Fig. 10.   Correlation of experimental and predicted kinetic parameters with testing and training data sets (a,c,e), as well as validation data set (b,d,f) for butt segment of 

asparagus during thermal treatments using the optimal ANN.

 Table 7

Errors in the prediction of  k , t 1/2  and  D-value with the optimal ANN for different segments of asparagus during thermal treatments.

MSE MAE MRE (%) STD A   STDR  (%)

k t 1/2   D-value   k (min1)   t 1/2  (min)   D-value (min)   k t 1/2   D-value

Bud segment 1.3925 0.5283 0.1309 0.0846 0.1307 0.0271 0.8301 1.8623 0.2224 0.1456 0.2255

Upper segment 2.4618 0.6436 0.0419 0.0519 0.0099 0.0033 2.6212 0.5899 0.0657 0.0706 0.0094

Middle segment 0.8985 0.4258 0.0394 0.0631 0.0163 0.0042 1.4624 0.7461 0.0682 0.1093 0.0282

Butt segment 0.2707 0.1883 0.0032 0.0165 0.0041 0.0004 0.8965 0.0670 0.0051 0.0269 0.0051

5600   H. Zheng et al. / Expert Systems with Applications 38 (2011) 5591–5602

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 11/12

& Domingo, 2004), shelf life (Siripatrawan & Jantawat, 2008),microbiology (Garcıa-Gimeno, Hervás-Martınez, & de Sióniz,

2002), drying applications (Kaminski & Tomczak, 2000;Kerdpiboon,Kerr, & Devahastin, 2006; Tripathy & Kumar, 2009), thermal process(Chen & Ramaswamy, 2003; Gonçalves, Minim, Coimbra, & Minim,2005), food freezing (Goñi, Oddone, Segura, Mascheroni, &

Salvadori, 2008; Mittal & Zhang, 2000). ANN is a powerful

mathematical tool that is capable of approximating any underlyingrelationship between the dependent and independent variables,learning from examples through iteration, without requiring a prior

knowledge of the relationships of the process parameters. More-over, its structure is relatively simple, with connections in paralleland sequence between neurons, which means a short computingtime and a high potential of robustness and adaptive performance

(Palancar et al., 1998). In this study, we tried to develop amathematical model based on artificial neural network to predictkinetics of AA degradation in asparagus during thermal process.

Our results showed that an one-hidden-layer FFBP has been builtable to predict the kineticparameters (k, t 1/2 and D-value) of AA loss,and the optimal number of nodes in the hidden layer was 24, 26, 26and 18 for bud, upper, middle and butt segments of asparagus,

respectively (Fig. 6). The optimal ANN could predict the kineticparameters of AA degradation in asparagus with an MSE of 1.3925and MAE 0.5283 for bud segment, MSE 2.4618 and MAE 0.6436for upper segment, MSE 0.8985 and 0.4258 for middle segment

and MSE 0.2707 and MAE 0.1883 for butt segment, as shown byTable 7. In addition, the correlation coefficients between experi-mental  k,   t 1/2  or  D-value and ANN predicted values were greaterthan 0.99 in all cases (Figs. 7–10). Therefore, ANN algorithm

provided dramatically low prediction error and gave high determi-nation coefficient indicating a very good fit between actual andpredicted kinetic parameters. ANN offers several advantages overtraditional digital computations, including faster speed of informa-

tion processing, learning ability, fault tolerance, and multi-outputability. Success of this research will provide chemical kinetics withan alternative method for kinetic parameters determination.

References

Afaghi, M., Ramaswamy, H. S., & Prasher, S. O. (2001). Thermal process calculationsusing artificial neural network models.  Food Research International, 34, 55–65.

Al-Zubaidy, M. M. I., & Khalil, R. A. (2007). Kinetic and prediction studies of ascorbicacid degradation in normal and concentrate local lemon juice during storage.Food Chemistry, 101, 254–259.

AOAC (2000).   Official method of analysis   (17th ed.). Gaithersburg, MD, USA:Association of Official Analytical Chemists [No. 967.21 Ascorbic acid invitamin preparation and juices].

Arroqui, C., Rumsey, T. R., Lopez, A., & Virseda, P. (2002). Losses by diffusion of ascorbic acid during recycled water blanching of potato tissue.  Journal of FoodEngineering, 52, 25–30.

Bila, S., Harkouss, Y., Ibrahim, M., Rousset, J., N’Goya, E., Baillargeat, D., et al. (1999).An accurate wavelet neural-network-based model for electromagneticoptimization of microwave circuits.   International Journal of RF and Microwave

Computer-Aided Engineering, 93, 297–306.Bucinski, A., Zielinski, H., & Kozłowska, H. (2004). Artificial neural networks forprediction of antioxidant capacity of cruciferous sprouts.  Trends in Food Scienceand Technology, 15, 161–169.

Burdurlu, H. S., Koca, N., & Karadeniz, F. (2006). Degradation of vitamin C in citrus juice concentrates during storage. Journal of Food Engineering, 74, 211–216.

Cabrera, A. C., & Prieto, J. M. (2010). Application of artificial neural networks to theprediction of the antioxidant activity of essential oils in two experimentalin vitro models.  Food Chemistry, 118, 141–146.

Chen, A., Leung, M. T., & Hazem, D. (2003). Application of neural networks to anemerging financial market: Forecasting and trading the Taiwan Stock Index.Computers and Operations Research, 30, 901–923.

Chen, C. R., & Ramaswamy, H. S. (2003). Analysis of critical control points in deviantthermal processes using artificial neural networks.  Journal of Food Engineering,57 , 225–235.

Cybenco, G. (1989). Approximation by superposition of a sigmoidal function.Mathematics of Control, Signals and Systems, 2 , 303–314.

Davies, M. B., Austin, J., & Partridge, D. A. (1991).   Vitamin C: Its chemistry andbiochemistry. Cambridge: Royal Society of Chemistry.

Dogan, A., Demirpence, H., & Cobaner, M. (2008). Prediction of groundwater levelsfrom lake levels and climate data using ANN approach.  Water SA, 34(2), 1–10.

Frias, J. M., & Oliveira, J. C. (2001). Kinetic models of ascorbic acid thermaldegradation during hot air drying of maltodextrin solutions.   Journal of FoodEngineering, 47 , 255–262.

Frias, J. M., Oliveira, J. C., Cunha, L. M., & Oliveira, F. A. (1998). Application of D-optimal design for determination of the influence of water content on thethermal degradation kinetics of ascorbic acid at low water contents.  Journal of Food Engineering, 38, 69–85.

Garcıa-Gimeno, R. M., Hervás-Martınez, C., & de Sióniz, M. I. (2002). Improvingartificial neural networks with a pruning methodology and genetic algorithmsfor their application in microbial growth prediction in food.  International Journal

of Food Microbiology, 72, 19–30.Garrote, R. L., Silva, R., & Bertone, R. A. (1986). Losses by diffusion of ascorbic acid

during water blanching of potato tissue.   Lebensmittel-Wissensenschaft undTechnologie, 19, 263–265.

Geeraerd, A. H., Herremans, C. H., Cenens, C., & Van Impe, J. F. (1998). Application of artificial neural networks as a non-linear modular modeling technique todescribe bacterial growth in chilled food products.  International Journal of FoodMicrobiology, 44, 49–68.

Gestal, M., Gómez-Carracedo, M. P., Andrade, J. M., Dorado, J., Fernández, E., Prada,D., et al. (2004). Classification of apple beverages using artificial neuralnetworks with previous variable selection.   Analytica Chimica Acta, 524,225–234.

Giannakourou, M. C., & Taoukis, P. S. (2003). Kinetic modelling of vitamin C loss infrozen green vegetables under variable storage conditions.  Food Chemistry, 83,33–41.

Gonçalves, E. C., Minim, L. A., Coimbra, J. S. R., & Minim, V. P. R. (2005). Modelingsterilization process of canned foods using artificial neural networks.  ChemicalEngineering and Processing, 44, 1269–1276.

Goñi, S. M., Oddone, S., Segura, J. A., Mascheroni, R. H., & Salvadori, V. O. (2008).Prediction of foods freezing and thawing times: Artificial neural networks andgenetic algorithm approach.  Journal of Food Engineering, 84, 164–178.

Guyer, D., & Yang, X. (2000). Use of genetic artificial neural networks and spectralimaging for defect detection on cherries.   Computers and Electronics in

 Agriculture, 29(3), 179–194.Hagan, M. T., & Menhaj, M. B. (1994). Training feed forward networks with the

Marquardt algorithm.  IEEE Transactions on Neural Networks, 6 , 861–867.Haralampu, S. G., & Karel, M. (1983). Kinetic models for moisture dependence of 

ascorbic acid and b-carotene degradation in dehydrated sweet potato.  Journal of Food Science, 48, 1872–1873.

Haykin, S. (1998).   Neural networks – A comprehensive foundation   (2nd ed.). UpperSaddle River, NJ: Prentice-Hall. pp. 26–32.

Hornik, K., Stinchcombe, M., & White, H. (1989). Multilayer feed forward networksare universal approximators.  Neural Networks, 2, 359–366.

Huelin, F. E. (1953). Studies on the anaerobic decomposition of ascorbic acid.  FoodResearch, 18, 633–639.

 Johnson, J. R., Braddock, R. J., & Chen, C. S. (1995). Kinetics of ascorbic acid loss andnonenzymatic browning in orange juice serum: Experimental rate constants.

 Journal of Food Science, 60(3), 502–505.Kaminski, W., & Tomczak, E. (2000). Degradation of ascorbic acid in drying process –

A comparison of description methods.  Drying Technology, 18, 777–790.Kerdpiboon, S., Kerr, W. L., & Devahastin, S. (2006). Neural network prediction of 

physical property changes of dried carrot as a function of fractal dimension andmoisture content.  Food Research International, 39, 1110–1118.

Lee, H. S., & Coates, G. A. (1999). Vitamin C in frozen, fresh squeezed, unpasteurized,polyethylene-bottled orange juice: A storage study.   Food Chemistry, 65,165–168.

Levenberg, K. (1944). A method for the solution of certain non-linear problems inleast squares.  Quarterly Journal of Applied Mathematics, 2(2), 164–168.

Lewis, E., Sheridan, C., O’Farrell, M., Flanagan, C., Kerry, J. F., & Jackman, N. (2008).Optical fibre sensors for assessing food quality in full scale production ovens – Aprincipal component analysis and artificial neural network based approach.Nonlinear Analysis: Hybrid Systems, 2, 51–57.

Marquardt, D. W. (1963). An algorithm for least-squares estimation of non-linearparameters. Journal of the Society for Industrial and Applied Mathematics, 2(2),431–441.

Mishkin, M., Saguy, I., & Karel, M. (1984). Optimization of nutrient during

processing: Ascorbic acid in potato dehydration.   Journal of Food Science, 49,1262–1265.

Mittal, G. S., & Zhang, J. (2000). Prediction of freezing time for food products using aneural network.  Food Research International, 33, 557–562.

Nadal, M., Espinosa, G., Schuhmacher, M., & Domingo, J. L. (2004). Patterns of PCDDsand PCDFs in human milk and food and their characterization by artificialneural networks.  Chemosphere, 54, 1375–1382.

Ni, H., & Gunasekaran, S. (1998). Food quality prediction with neural networks.  FoodTechnology, 52(10), 60–65.

Nindo, C. I., Sun, T., Wang, S. W., Tang, J., & Powers, J. R. (2003). Evaluation of dryingtechnologies for retention of physical quality and antioxidants in asparagus( Asparagus officinalis   L.).   Lebensmittel-Wissensenschaft und Technologie, 36 ,507–516.

O’Farrell, M., Lewis, E., Flanagan, C., Lyons, W. B., & Jackman, N. (2005). Combiningprincipal component analysis with an artificial neural network to performonline quality assessment of food as it cooks in a large-scale industrial oven.Sensors and Actuators B, 107 , 104–112.

Palancar, M. C., Aragón, J. M., & Torrecilla, J. S. (1998). pH-control system based on

artificial neural networks.   Industrial and Engineering Chemistry Research, 37 (7),2729–2740.

H. Zheng et al. / Expert Systems with Applications 38 (2011) 5591–5602   5601

8/13/2019 Degradacion Ac. Ascorbico

http://slidepdf.com/reader/full/degradacion-ac-ascorbico 12/12

Panigrahi, S., Balasubramanian, S., Gu, H., Logue, C., & Marchello, M. (2006). Neural-network-integrated electronic nose system for identification of spoiled beef.LWT, 39, 135–145.

Ruan, R., Almaer, S., & Zhang, J. (1995). Prediction of dough rheological propertiesusing neural networks.  Cereal Chemistry, 72(3), 308–311.

Shao, Y., Poobrasert, O., Kennelly, E. J., Chin, C. K., Ho, C. T., Huang, M. T., et al. (1997).Steroidal saponins from Asparagus officinalis and their cytotoxic activity.   PlantaMedica, 63(3), 258–262.

Siripatrawan, U., & Jantawat, P. (2008). A novel method for shelf life prediction of apackaged moisture sensitive snack using multilayer perceptron neural network.

Expert Systems with Applications, 34, 1562–1567.Tripathy, P. P., & Kumar, S. (2009). Neural network approach for food temperature

prediction during solar drying.   International Journal of Thermal Sciences, 48,1452–1459.

Uddin, M. S., Hawlader, M. N. A., Ding, L., & Mujumdar, A. S. (2002). Degradation of ascorbic acid in dried guava during storage.   Journal of Food Engineering, 51,21–26.

Vieira, M. C., Teixeira, A. A., & Silva, C. L. M. (2000). Mathematical modeling of thethermal degradation kinetics of vitamin C in cupuaçu (Theobroma grandiflorum)nectar. Journal of Food Engineering, 43, 1–7.

Villota, R., & Karel, M. (1980). Prediction of ascorbic acid retention during drying. II.Simulation of retention in a model system.   Journal of Food Processing andPreservation, 4, 141–159.

Vinson, J. A., Hao, Y., Su, X. H., & Zubik, L. (1998). Phenol antioxidant quantity andquality in foods: Vegetables.   Journal of Agricultural and Food Chemistry, 46 (9),3630–3634.

Zhang, Y., Hu, X. S., Chen, F., Wu, J. H., Liao, X. J., & Wang, Z. F. (2008). Stability and

color characteristics of PEF-treated cyanidin-3-glucoside during storage.   FoodChemistry, 106 , 669–676.

Zhang, G. P., & Qi, G. M. (2005). Neural network forecasting for seasonal and trendtime series.  European Journal of Operational Research, 160, 501–514.

5602   H. Zheng et al. / Expert Systems with Applications 38 (2011) 5591–5602