پیش بینی محصولات بنیادی روغن بابونه Abdicate Matricaria chamomilla L. با استفاده از سیستم شبکه عصبی مصنوعی
محورهای موضوعی : مجله گیاهان زینتینازنین خاکی پور 1 * , مهتاب پاینده 2
1 - دانشکده علوم خاکشناسی، واحد سوادکوه، دانشگاه آزاد اسلامی، سوادکوه، ایران
2 - دانشجوی کارشناسی ارشد علوم باغبانی، واحد علوم و تحقیقات، دانشگاه آزاد اسلامی، تهران، ایران
کلید واژه: نیتروژن, شبکه عصبی مصنوعی (ANN), معادل کربنات کلسیم (CCE), پرسپترون چند لایه,
چکیده مقاله :
هدف از این تحقیق، پیش‌بینی نسبت و تولید اسانس بابونه با استفاده از یک سیستم شبکه عصبی مصنوعی متکی بر ویژگی‌های فیزیکوشیمیایی خاص خاک بود. سایت های مختلف کشت بابونه مورد بررسی قرار گرفت و 100 نمونه خاک به گلخانه منتقل شد. pH، EC، K، OM (ماده آلی)، CCE (معادل کربنات کلسیم) و میزان رس در خاک ها از 8.75 تا 7.94، 1.6 تا 1.0، 381 تا 135، 2.30 تا 0.22، 69 تا 16، و 6.5 متغیر بود. به ترتیب به 32.0 رسید. پارامترهای رشد، درصد اسانس و عملکرد اندازه گیری شد. مدل‌سازی شبکه عصبی مصنوعی با هدف پیش‌بینی غلظت و عملکرد اسانس با استفاده از سه مجموعه از ویژگی‌های خاک به عنوان پیش‌بینی‌کننده انجام شد: نیتروژن (N)، فسفر (P)، پتاسیم (K)، و خاک رس. pH، EC، مواد آلی (OM) و خاک رس. CCE، خاک رس، سیلت، ماسه، N، P، K، OM، pH و EC. در نتیجه، سه تابع انتقال (PTF) با استفاده از پرسپترون چند لایه (MLP) با الگوریتم آموزشی Levenberg-Marquardt برای تخمین محتوای اسانس بابونه فرموله شد. ارزیابی نتایج نشان داد که PTF سوم (PTF3) که با استفاده از تمامی متغیرهای مستقل توسعه یافته است، بالاترین دقت و پایایی را از خود نشان می دهد. علاوه بر این، یافته‌ها امکان پیش‌بینی غلظت و عملکرد اسانس بابونه را بر اساس ویژگی‌های فیزیکوشیمیایی خاک پیشنهاد کرد. این پیامدهای قابل توجهی برای ارزیابی تناسب زمین، شناسایی مناطق مساعد برای کشت بابونه و برنامه ریزی برای بازده اسانس دارد.
The aim of this research was to forecast the proportion and production of chamomile essential oils by employing an artificial neural network system reliant on specific soil physicochemical characteristics. Various chamomile cultivation sites were explored, and 100 soil samples were transported to the greenhouse. The pH, EC, K, OM (organic matter), CCE (calcium carbonate equivalent), and clay content in the soils ranged from 8.75 to 7.94, 1.6 to 1.0, 381 to 135, 2.30 to 0.22, 69 to 16, and 55.6 to 32.0, respectively. Growth parameters, essential oil percentage, and yield were measured. The artificial neural network modeling aimed to predict essential oil concentration and yield using three sets of soil properties as predictors: Nitrogen (N), phosphorus (P), potassium (K), and clay; pH, EC, organic matter (OM), and clay; CCE, clay, silt, sand, N, P, K, OM, pH, and EC. Consequently, three pedotransfer functions (PTFs) were formulated using the multi-layer perceptron (MLP) with the Levenberg-Marquardt training algorithm to estimate chamomile essential oil content. The evaluation of results indicated that the third PTF (PTF3), developed using all independent variables, exhibited the highest accuracy and reliability. Furthermore, the findings suggested the feasibility of predicting chamomile essential oil concentration and yield based on soil physicochemical properties. This has significant implications for land suitability assessments, identifying areas conducive to chamomile cultivation, and planning for essential oil yields.
Ajili Lahiji, A., Mohammadi Torkashvand, A., Mehnatkesh, A. and Navidi, M. 2018. Status of macro and micro nutrients of olive orchard in northern Iran. Asian Journal of Water, Environment and Pollution, 15(4): 143-148.
Akbar, A., Kuanar, A., Patnaik, J., Mishra, A. and Nayak, S. 2018. Application of artificial neural network modeling for optimization and prediction of essential oil yield in turmeric (Curcuma longa L.). Computers and Electronics in Agriculture, 48: 160-178.
Belal, B.E.A., El-Kenawy, M.A. and Uwakiem, M.K. 2016. Foliar application of some amino acids and vitamins to improve growth, physical and chemical properties of flame seedless grapevines. Egyptian Journal of Horticulture, 43: 123-136.
Benier, G., Kinet, J.M. and Sachs, R.M. 1985. Handbook of flowering, ed. Halevy, H. CRC press. pp. 423-434.
Bennett, W. 1993. Plant nutrient utilization and diagnostic plant symptoms. In: W.F. Bennett (Ed). Nutrient deficiencies and toxicities in crop plants. APS Press, St Paul, Minnesota.
Besalatpour, A.A., Ayoubi, S., Hajabbasi, M.A., Mosaddeghi, M.R. and Schulin, R. 2013. Estimating wet soil aggregate stability from easily available properties in a highly mountainous watershed. Catena, 111: 72-79.
Bocco, M., Willington, E. and Arias, M. 2010. Comparison of regression and neural networks models to estimate solar radiation. Chilean Journal of Agricultural Research, 70: 428-435.
Cecílio Filho, L.A.B., Feltrim, A.L., Mendoza Cortez, J.W., Gonsalves, M.V., Pavani, L.C. and Barbosa, J.C. 2015. Nitrogen and potassium application by fertigation at different watermelon planting densities. Journal of Soil Science and Plant Nutrition, 15 (4): 928-937.
Chrysargyris, A., Xylia, P., Botsaris, G. and Tzortzakis, N. 2017. Antioxidant and antibacterial activities, mineral and essential oil composition of spearmint (Mentha spicata L.) affected by the potassium levels. Industrial Crops and Products, 103: 202-212.
Dai, F., Zhou, Q., Lv, Z., Wang, X. and Liu, G. 2014. Spatial prediction of soil organic matter content integrating artificial neural network and ordinary kriging in Tibetan Plateau. Ecological Indicators, 45: 184-194.
Dai, L., Li, Y., Wang, Q., Niu, J., Li, W., Wang, Y., Wang, J., Zhou, Z. and Gao, N. 2022. Chamomile: A review of its traditional uses, chemical constituents, pharmacological activities and quality control studies. Molecules, 28(1): 133. https://doi.org/10.3390/molecules28010133.
Emami, A.S. 1996. Methods of plant analysis, Vol. II. Technical Journal No. 982. Karaj, Iran: Soil and Water Research Institute (In Persian).
Goos, R.G. 1995. A laboratory exercise to demonstrate nitrogen mineralization and immobilization. Journal of Natural Resources and Life Sciences Education, 24: 68-70.
Jat, R.S. and Ahaheat, I.P.S. 2006. Direct and residual effects of vermicomposts, biofertilizer and phosphorus on soil nutrient dynamics and productivity of chickpea-fodder maize sequence. Journal of Sustainable Agriculture, 28(1): 41-54.
Klute, A. 1986. Methods of soil analysis. Part 1: Physical and mineralogical methods. 2nd Edition, American Society of Agronomy and Soil Science Society of America, Madison, USA.
Malakouti, M.J., Karimian, N. and Keshavarz, P. 2008. Determination methods for nutritional deficiencies and recommendations for fertilizer. Office for the Publishing of Scientific Works. Tehran, Iran. (In Persian)
Malakouti, M. J. and Shahabi, A. 2000. Chloro-calcium spraying is an indispensable necessity in improving the qualitative characteristics of apple. Deputy Directorate for the Promotion of TAT. Ministry of Agriculture, Tehran. (In Persian)
Marashi, M., Mohammadi Torkashvand, A., Ahmadi, A. and Esfandyari, M. 2017. Estimation of soil aggregate stability indices using artificial neural network and multiple linear regression models. Spanish Journal of Soil Science, 7: 122-132.
Marashi, M., Mohammadi Torkashvand, A., Ahmadi, A. and Esfandyari, M. 2019. Adaptive neuro-fuzzy inference system: Estimation of soil aggregates stability. Acta Ecologica Sinica, 39: 95-101.
Moghimi, S., Mahdian, M.H., Parvizi, Y. and Masihabadi, M.H. 2014. Estimating effects of terrain attributes on local soil organic carbon content in a semi-arid pastureland. Journal of Biodiversity and Environmental Sciences, 5(2): 67-106.
Mohammadi Torkashvand, A., Ahmadi, A., Gómez, P.A. and Maghoumi, M. 2019. Using artificial neural network in determining postharvest life of kiwifruit. Journal of The Science of Food and Agriculture, 99(13): 5918-5925. https://doi.org/10.1002/jsfa.9866
Mohammadi Torkashvand, A., Ahmadi, A. and Nikravesh, N. 2017. Prediction of kiwifruit firmness using fruit mineral nutrient concentration by artificial neural network (ANN) and multiple linear regressions (MLR). Journal of Integrative Agriculture, 16(7): 1634-1644.
Mohammadi Torkashvand, A., Ahmadipour, A. and Khaneghah, A.M. 2020. Estimation of kiwifruit yield by leaf nutrients concentration and artificial neural network. The Journal of Agricultural Science, 158 (3): 185-193.
Mokhtari Karchegani, P., Ayoubi, Sh., Honarju, N. and Jalalian, A. 2011. Predicting soil organic matter by artificial neural network in landscape scale using remotely sensed data and topographic attributes. Geophysical Research Abstracts, 13, article no. EGU2011-1075. Available online from https://meetingorganizer.copernicus.org/EGU2011/EGU2011-1075.pdf
Niazian, M., Sadat Noori, S. and Abdipour, M. 2018. Artificial neural network and multiple regression analysis models to predict essential oil content of ajowan (Carum copticum L.). 9: 124-131.
Nurzynska-Wierdak, R. 2013. Does mineral fertilization modify essential oil content and chemical composition in medicinal plants? Acta Scientiarum Polonorum Hortorum Cultus, 12: 3-16.
Pacheco, C., Calouro, F., Vieira, S., Santos, F., Neves, N., Curado, F., Franco, J., Rodrigues, S. and Antunes, D. 2008. Influence of nitrogen and potassium on yield, fruit quality and mineral composition of kiwifruit. International Journal of Energy and Environment, 2: 9-15.
Paye, A.L., Miller, R.H. and Keeney, D.R. 1948. Method of soil analysis, Part II., SSSA Inc.
Peng, J., Zhang, Y.Z. and Pang, X.A. 2012. Hyperspectral features of soil organic matter content in South Xinjiang. Arid Land Geography, 55(5): 740–746.
Petronilho, S., Maraschin, M., Delgadillo, I., Coimbra, M.A. and Rocha, S.M. 2011. Sesquiterpenic composition of the inflorescences of Brazilian chamomile (Matricaria recutita L.): Impact of the agricultural practices. Industrial Crops and Products, 34: 1482–1490. doi: 10.1016/j.indcrop.2011.05.005
Poorghadir, M., Mohammadi Torkashvand, A., Mirjalili, S.A. and Moradi, P. 2021. Interactions of amino acids (proline and phenylalanine) and biostimulants (salicylic acid and chitosan) on the growth and essential oil components of savory (Satureja hortensis L.). Biocatalysis and Agricultural Biotechnology, 30: 101815.
Radkowski, A. and Radkowska, I. 2018. Influence of foliar fertilization with amino acid preparations on morphological traits and seed yield of timothy. Plant, Soil and Environment, 64: 209-213.
Rahmani Khalili, M., Asadi, M.E., Mohammadi Torkashvand, A. and Pazira, E. 2020. Regression analysis for yield comparison of saffron as affected by physicochemical properties of the soil, case study in northeast of Iran. Agricultural Research, 9(12): 568-76. https://doi.org/10.1007/s40003-020-00455-6
Savitikadi, P., Jogam, P., Khan Rohela, G., Ellendula, R., Sandhya, D., Rao Allini, V. and Abbagani, S. 2020. Direct regeneration and genetic fidelity analysis of regenerated plants of Andrographis echioides (L.) - An important medicinal plant. Industrial Crops and Products, 155: 112766.
Soltanpour, P.N. and Schwab, A.P. 1977. A new soil test for simultaneous extraction of macro- and micronutrients in alkaline soils. Communication in Soil Science and Plant Analysis, 8: 195-207.
Tashakori, F., Mohammadi Torkashvand, A., Ahmadi, A. and Esfandiari, M. 2020. Comparison of different metthods of estimating saffron yield based on soil properties in Golestan province. Communications in Soil Science and Plant Analysis, 51(13): 1767-1779.
Tofighi Alikhani, T., Tabatabaei, S.J., Mohammadi Torkashvand, A., Khalighi, A. and Talei, D. 2021. Effects of silica nanoparticles and calcium chelate on the morphological, physiological and biochemical characteristics of gerbera (Gerbera jamesonii L.) under hydroponic condition. Journal of Plant Nutrition, 44(7): 1039-1053.
Ubessi, C., Tedesco, S.B., de Bona da Silva, C., Baldoni, M., Krysczun, D.K., Heinzmann, B.M., Rosa, I.A. and Mori, N.C. 2019. Antiproliferative potential and phenolic compounds of infusions and essential oil of chamomile cultivated with homeopathy. Journal of Ethnopharmacology, 239: 111907. doi: 10.1016/j.jep.2019.111907
Upadhyay, R. K. and Patra, D.D. 2011. Influence of secondary plant nutrient (Ca and Mg) on growth and yield of chamomile (Matricaria recutita L.). Asian Journal of Crop Science, 3: 151-157.
Walkley, A.J. and Black, I.A. 1934. Estimation of soil organic carbon by the chromic acid titration method. Soil Science, 37: 29-38.
Wan, W.T., Song, Y.J., Xu, L.J., Xiao, P.G. and Miao, J.H. 2019. Research review and application prospect analysis of matricaria. Modern Chinese Medicine, 21: 260–265. doi: 10.13313/j.issn.1673-4890.20181008002
Zhao, Y.F. 2018. Study on chemical composition and quality standard of Uyghur chamomile. Master’s Thesis. China Academy of Chinese Medical Sciences; Beijing, China.
Zhao, D.S., Han, S.L., Yi, Y.J., Qiu, L., Ren, L.J. and Li, X.X. 2015. Standard operating procedure for standardized planting of local medicinal herb German chamomile. Journal of Anhui Agricultural Sciences, 43: doi: 10.13989/j.cnki.0517-6611.2015.28.028
Zhou, T., Shi, P.J., Luo, J.Y. and Shao, Z.Y. 2008. Estimation of soil organic carbon based on remote sensing and process model. Frontiers of Forestry in China, 3: 139–147.
Expectation of chamomile fundamental oil abdicate by using the Artificial Neural Network System
Nazanin Khakipour1* and Mahtab Payandeh2
1-Department of Soil Science, Savadkooh Branch, Islamic Azad University, Savadkooh, Iran
2-Department of Horticulture, Science and Research Branch, Islamic Azad University, Tehran, Iran
*Corresponding author: nazanin_kh_43713@yahoo.com
ABSTRACT
The aim of this research was to forecast the proportion and production of chamomile essential oils by employing an artificial neural network system reliant on specific soil physicochemical characteristics. Various chamomile cultivation sites were explored, and 100 soil samples were transported to the greenhouse. The pH, EC, K, OM (organic matter), CCE (calcium carbonate equivalent), and clay content in the soils ranged from 8.75 to 7.94, 1.6 to 1.0, 381 to 135, 2.30 to 0.22, 69 to 16, and 55.6 to 32.0, respectively. Growth parameters, essential oil percentage, and yield were measured. The artificial neural network modeling aimed to predict essential oil concentration and yield using three sets of soil properties as predictors: nitrogen (N), phosphorus (P), potassium (K), and clay; pH, EC, organic matter (OM), and clay; CCE, clay, silt, sand, N, P, K, OM, pH, and EC. Consequently, three pedotransfer functions (PTFs) were formulated using the multi-layer perceptron (MLP) with the Levenberg-Marquardt training algorithm to estimate chamomile essential oil content. The evaluation of results indicated that the third PTF (PTF3), developed using all independent variables, exhibited the highest accuracy and reliability. Furthermore, the findings suggested the feasibility of predicting chamomile essential oil concentration and yield based on soil physicochemical properties. This has significant implications for land suitability assessments, identifying areas conducive to chamomile cultivation, and planning for essential oil yields.
Keywords: Artificial Neural Network (ANN), calcium carbonate equivalent (CCE), multilayer perceptron, nitrogen.
INTRODUCTION
Ensuring optimal nutrition is a crucial factor in enhancing both the quality and quantity of crops, a condition profoundly impacted by the soil environment (Ajili et al., 2018; Tofighi Alikhani, 2021). The role of soil characteristics in influencing crop yield has been emphasized by Peng et al. (2012), highlighting the indispensability of mineral elements in the process of photosynthesis for the production of organic materials in leaves. Each macronutrient plays a distinct role in the metabolic processes governing plant growth (Benier, 1985). The nutritional status, particularly the balance between elements extracted from air and soil, significantly influences plant flowering.
Given the significance of advancing the cultivation of medicinal plants and harnessing their products as natural, health-compatible ingredients, employing diverse cultivation and nutrition methods is essential to augment the essential oil and active compound content of medicinal plants (Lahiji et al., 2019; Savitikadi et al., 2020). Chamomile is an annual or perennial plant belonging to the Asteraceae family. The plant improves appetite and reduces swelling, pain and sweating(Ubessi et al., 2019). Chamomile are native to temperate regions of Asia and Europe and are grown around the world for their high medicinal, cosmetic, and dietary values(Wan et al., 2019). It has been used for thousands of years in ancient Greece, Rome and Egypt. In China, the detailed uses of this plant were first recorded in Uyghur medicine. Experienced TCM practitioners believe that preparations containing chamomile have a soothing effect. Additionally, the plant is used in traditional, homeopathic and Unani preparations(Zhao et al., 2015). There are two main varieties of Chamomile: Matricaria chamomilia L. and Anthemis nobilis (L.) Matricaria chamomilia L. belongs to the genus Matricaria. This is an annual plant and the flowering period is from May to July in China. Chamaemelum nobile L. is a perennial plant of the genus Chamaemelum. The flowering period is from April to May in China (Fen et al., 2021) Matricaria chamomilia L. is relatively common and has been the subject of much research and use. Currently, 26 countries around the world have included this plant in their pharmacopoeia. Chamomile heads are often used as medicine(Petronilho et al., 2011). Chamomile contains flavonoids, coumarins, volatile oils, terpenes, sterols, organic acids and polysaccharides, among other compounds. Possessing a wide range of compounds, Chamomile has various pharmacological activities such as anti-cancer, anti-infectious, anti-inflammatory, anti-oxidant, hypoglycemic, hypotensive, hypolipidemic, anti-allergic, anti-depressant and neuroprotective, among many other effects(Zhao et al., 2018). In general, this plant has special research value. However, there is little criticism on this topic in the literature. This article provides a comprehensive overview of plant properties and distribution, traditional uses, chemical composition, pharmacological effects, and quality control methods of chamomile(Dai et al., 2022).
Notably, the growth and essential oil yield of chamomile are influenced by factors such as calcium and magnesium, with magnesium exhibiting a more significant impact than calcium (Upadhyay and Patra, 2011). In the realm of agriculture, a key challenge lies in predicting crop yield based on readily accessible indicators. Various factors, including soil nutrition and physicochemical properties, affect plant yield and essential oil content (Belal et al., 2016; Radkowski and Radkowska, 2018; Mohammadi Torkashvand et al., 2020). Hypotheses suggest that predicting yield is feasible by assessing nutrient concentrations in leaves (Mohammadi Ajili et al., 2018; Mohammadi Torkashvand et al., 2020), fruits (Mohammadi Torkashvand et al., 2019), or soil characteristics (Rahmani Khalili et al., 2020; Tashakori et al., 2020). Such predictions enable effective planning for fertilization, soil selection, and future financial considerations for farmers.
In predicting natural variables, transfer functions, especially artificial neural networks (ANN), have proven more efficient in heterogeneous natural systems like soil and plants compared to traditional regression methods. Numerous studies have successfully employed ANN to estimate soil variables and predict crop yield based on factors such as weather, soil properties, and growth characteristics (Zhou et al., 2008; Bocco et al., 2010; Mokhtari Karchegani et al., 2011; Besalatpour et al., 2013; Dai et al., 2014; Moghimi et al., 2014; Marashi et al., 2017; Marashi et al., 2019). Tashakori et al. (2020) specifically highlighted the superior accuracy of ANN in estimating saffron yield compared to multiple linear regression and adaptive neuro-fuzzy inference system models.
A study by Poorghadir et al. (2021) underscored the influence of soil properties and nutrition on the yield and essence percentage of crops. This study aims to delve into the impact of soil characteristics on the concentration and quantity of chamomile essential oils, exploring the feasibility of using artificial neural networks to estimate these essential oil parameters based on critical physicochemical soil properties.
MATERIALS AND METHODS
Soil experiments
Multiple locations for Chamomile cultivation in Kermanshah and Hamadan provinces, West of Iran, were investigated. From 20 areas, a total of 100 soil samples (five from each area) were collected at a depth of 0-30 cm and sent to IAU, Science and Research Branch, Tehran, Iran. The sampling areas shared similar topographic and climatic characteristics. Soil samples underwent air-drying, and clods were fragmented into small particles using a plastic hammer before passing through a 2-millimeter sieve (Klute, 1986). Laboratory analyses, focusing on phosphorus, nitrogen, potassium nutrients, pH, Electrical Conductivity (EC), texture, and organic matter, were conducted using 0.5 kg of each soil sample. The remaining soil was dedicated to greenhouse experiments. Soil pH and EC were measured in saturated soil extracts, soil texture was determined hydrometrically, and calcium carbonate equivalent (CCE) was assessed through titration (Paye et al., 1948). Nitrogen was measured via the Kjeldahl method (Goos, 1995), and available potassium and phosphorus concentrations were determined by flame emission and spectrophotometry methods after soil extraction (Soltanpour and Schwab, 1977; Emami, 1996). Organic matter content was determined by the Walkley and Black method (1934). Statistical data regarding soil properties are presented in Table 1.
Greenhouse experiment
In a completely randomized design, 100 distinct soil samples were applied to plots with dimensions of 30 by 35 cm and a depth of 25 cm, with 20 seeds planted in each plot. Following germination and early growth, the number of plants per plot was reduced to 10 during the quadruple stage. Standard field operations, including irrigation, weed control, and pest management, were uniform across all plots. Upon full flowering, flowers were harvested at a maximum length of five centimeters, dried at 60ºC, and the dry flower yield per plot (kg ha⁻¹) and essential oil concentration were determined. Essential oil content was measured using the Kelevenger apparatus and expressed as g/100 g of dry flowers, while essential oil yield was reported as kg ha⁻¹ based on dry flower yield.
Artificial Neural Network
The multilayer perceptron (MLP) learning rule, a feedforward network, was employed in this study. The Levenberg-Marquart back propagation algorithm facilitated training, utilizing the Tangent axon function as the activation function. The design of the artificial neural network was executed using NeuroSolutions 5.05 software. Data were divided into training (60%), validation (20%), and test (20%) sets. Three sub-series of variables were used as input variables for estimating essential oil percentage and yield. The efficiency of three pedotransfer functions (PTF1, PTF2, and PTF3) was compared, each developed using different sets of predictors. The evaluation criteria included the coefficient of determination (R²) and root mean square error (RMSE). The number of neurons in the input layer corresponded to the input parameters, while the complexity of the network was determined by the number of hidden layers. The output layer neurons matched the number of output parameters. Model precision and accuracy were assessed using the test set data after training with the training and validation set data.
(2)
In which: yi, 
, 
respectively, the measured dependent variable, its mean and the estimated dependent variable, and N is the number of observations. Other criteria used to evaluate the precision of transition functions were the Geometric Mean Error Ratio (GMER) and Geometric Standard Deviation of error ratio (GSDER):
The Geometric Mean of Error Ratio (GMER) serves as an indicator of the level of agreement between measured and estimated values. A GMER equal to one signifies a complete alignment between measured and predicted values. A GMER greater than one suggests that the predicted values surpass the measured values, while a GMER less than one indicates that the estimated values are lower than the measured values. On the other hand, the Geometric Standard Deviation of Error Ratio (GSDER) functions as a measure of data dispersion. A value close to one indicates minimal dispersion, and the deviation from one reflects the extent to which most estimates deviate from the measured data.
RESULTS AND DISCUSSION
Table 1. Correlation between the input variables of the neural network and the amount of essential oil and essential oil yield
Variable | pH | EC | N | P | K | OM | Lime | Sand | Silt | Clay | Essential oil | Essential oil Yield |
pH | 1 |
|
|
|
|
|
|
|
|
|
|
|
EC | 0.511** | 1 |
|
|
|
|
|
|
|
|
|
|
N | -0.101 | 0.152 | 1 |
|
|
|
|
|
|
|
|
|
P | 0.806** | 0.546** | -0.382* | 1 |
|
|
|
|
|
|
|
|
K | -0.390* | -0.419** | 0.099 | -0.323* | 1 |
|
|
|
|
|
|
|
OM | 0.414** | 0.675** | 0.362* | 0.443** | -0.046 | 1 |
|
|
|
|
|
|
Lime | -0.267 | 0.015 | 0.001 | -0.287 | 0.044 | -0.544** | 1 |
|
|
|
|
|
Sand | -0.032 | -0.339* | -0.439** | -0.029 | 0.268 | -0.577** | 0.648** | 1 |
|
|
|
|
Silt | -0.202 | 0.219 | 0.269 | -0.069 | -0.201 | 0.409** | -0.258 | -0.408** | 1 |
|
|
|
Clay | 0.188 | 0.178 | 0.243 | 0.084 | -0.12 | 0.276 | -.465** | -0.711** | -0.353* | 1 |
|
|
Essential oil | 0.092 | -0.024 | 0.391* | 0.278 | 0.434** | 0.355* | -0.181 | 0.155 | 0.153 | -0.277 | 1 |
|
Essential oil Yield | 0.235 | 0.163 | 0.401* | 0.423** | 0.382* | 0.449** | -0.101 | 0.226 | 0.246 | -0.421** | 0.919** | 1 |
Correlation among variables
The correlation between the variables examined and the percentage and yield of essential oils is depicted in Table 2, offering valuable insights into potential input data for the neural network. Our findings reveal a positive and significant correlation between the percentage and yield of essential oil with organic matter, N, P, K contents, and clay. Consistent with our results, Jat and Ahaheat (2006) demonstrated that bio-fertilizers containing nitrogen, phosphorus, and potassium enhance the growth and essential oil production of fennel plants. Phosphorus plays a pivotal role in seed and flower germination, vegetative growth, fruit maturation, and metabolic processes (Bennett, 1993; Malakouti and Shahabi, 2000; Malakouti et al., 2008). Additionally, potassium influences herbal essential oil quantity and quality by affecting metabolic pathways and enzymatic activity (Pacheco et al., 2008). The correlation findings align with previous studies emphasizing the role of these elements in enhancing essential oil production (Nurzynska-Wierdak, 2013; Cecílio Filho et al., 2015; Chrysargyris et al., 2017a).
Table 2. Input data for constructing a neural network in three different transfer functions and the characteristics of neural networks made.
Transition function | Model inputs | The number of hidden layers | Number of hidden layer nodes 1 | Number of hidden layer nodes 2 | Number of hidden layer nodes 3 | Type of transfer function | Type of target function |
PTF1 | N, P, K and clay | 3 | 4 | 2 | 1 | Tangent axon | Levenberg-Marquardt algorithm |
PTF2 | pH، EC، Organic matter and clay | 3 | 4 | 4 | 2 | Tangent axon | Levenberg-Marquardt algorithm |
PTF3 | All variables | 1 | 10 | - | - | Tangent axon | Levenberg-Marquardt algorithm |
Given the significant correlation with essential oils, three data series were designated as input for the artificial neural network (ANN), as outlined in Table 3, considering the number of hidden layers and nodes.
Table 3: Determination coefficient (R2), error (RMSE), GMER and GSDER in two sets of training and test data in predicting essential oil concentration (g/100 g)
Transition function | Data series | R2 | RMSE | GMER | GSDER |
PTF1 | Training | 0.7856 | 0.226 | 1.27 | 2.37 |
| test
| 0.8271 | 0.201 | 1.04 | 1.24 |
PTF2 | Training | 0.9662 | 0.072 | 0.91 | 2.27 |
| test
| 0.8954 | 0.237 | 0.83 | 1.57 |
PTF3 | training | 0.9562 | 0.023 | 0.98 | 1.22 |
| test | 0.9478 | 0.086 | 1.02 | 1.32 |
Estimation of essential oil
Figure 1 illustrates the distribution and alignment of actual (measured) and estimated essential oil percentages for three input data series. R² values, as presented in Table 4, indicate the accuracy of the ANN models in predicting essential oil percentages. Notably, the inclusion of all nine variables in the third transfer function (PTF3) resulted in a higher accuracy (R²) in both training (96.62%) and test (94.78%) data sets compared to PTF1 and PTF2. This increase in variables led to a reduction in error, as indicated by the GMER and GSDER values, demonstrating improved precision in the estimation of chamomile essential oil percentage.
Table 4: Determination coefficients (R2), error (RMSE), GMER and GSDER in two sets of training and test data to predict the essential oil yield (kg / ha)
Transition function | Data series | R2 | RMSE | GMER | GSDER |
PTF1 | training | 0.7984 | 1.27 | 1.47 | 2.16 |
| test
| 0.9125 | 0.670 | 1.01 | 1.25 |
PTF2 | training | 0.8646 | 0.536 | 1.19 | 1.38 |
| test
| 0.8323 | 0.998 | 1.26 | 1.45 |
PTF3 | training | 0.9654 | 0.088 | 0.98 | 1.10 |
| test | 0.9251 | 0.608 | 0.92 | 1.20 |
Fig. 1: Measured values (actual) and estimated concentration of essential oils in diagram 1: 1 test data and their conformance
Similarly, Figure 2 displays the dispersion of measured yield values and the estimated essential oil values for the test series. The accuracy of the model improved in the third transfer function (PTF3), with an R² of 91.25% in the test data set. Despite a lower estimation in PTF3, the model exhibited the least error, as indicated by the GMER and GSDER values, demonstrating enhanced accuracy in predicting chamomile essential oil yield.
Fig. 2: Measured (actual) and estimated essential oil yield in diagram 1: 1 test data and their conformance
These results align with findings from studies employing ANN for crop yield prediction, emphasizing the importance of considering multiple variables for improved accuracy (Mohammadi Torkashvand et al., 2017; Akbar et al., 2018; Niazian et al., 2018). In summary, the comprehensive consideration of all nine soil variables led to a more accurate and reliable estimation of chamomile essential oil content and yield, showcasing the potential of ANN models in precision agriculture and decision-making systems.
CONCLUSION
In summary, the findings indicate that the third transfer function, incorporating nine variables as input for the neural network, proved to be the most precise in estimating both the concentration and yield of essential oil. This function exhibited the highest R² values and the lowest RMSE values, showcasing superior accuracy. Furthermore, the estimated values from this function demonstrated the closest alignment with observed values, displaying minimal deviation from the 1:1 line. The proposed model's evaluation metrics for estimating essential oil percentage in the test series data were as follows: R² = 94.78, RMSE = 0.86, GMER = 1.02, and GSDER = 1.32. Similarly, for essential oil yield in the test series data, the metrics were R² = 91.51, RMSE = 0.608, GMER = 0.92, and GSDER = 1.20. These results affirm the high accuracy and precision of predicting chamomile essential oil concentration and yield based on soil physicochemical properties. This predictive capability holds significance in determining land suitability, allowing for the identification of areas conducive to chamomile cultivation and facilitating strategic planning for essential oil yields. Future research avenues could explore additional soil characteristics or combinations thereof in conjunction with artificial neural networks for chamomile essential oil prediction. Additionally, the evaluation of alternative models, such as neuro-fuzzy, warrants consideration for estimating the concentration and essential oil yields not only for chamomile but also for other medicinal plant species.
REFERENCES
Ajili Lahiji, A., Mohammadi Torkashvand, A., Mehnatkesh, A. and Navidi, M. 2018. Status of macro and micro nutrients of olive orchard in northern Iran. Asian Journal of Water, Environment and Pollution, 15(4): 143-148.
Akbar, A., Kuanar, A., Patnaik, J., Mishra, A. and Nayak, S. 2018. Application of Artificial Neural Network modeling for optimization and prediction of essential oil yield in turmeric (Curcuma longa L.). Computers and Electronics in Agriculture, 48: 160-178.
Belal, B.E.A., El-Kenawy, M.A. and Uwakiem, M.K. 2016. Foliar application of some amino acids and vitamins to improve growth, physical and chemical properties of flame seedless grapevines. Egyptian Journal of Horticulture, 43: 123-136.
Benier, G., Kinet, J.M. and Sachs, R.M. 1985. Handbook of flowering, ed. Halevy, H. CRC press. pp. 423-434.
Bennett, W. 1993. Plant nutrient utilization and diagnostic plant symptoms. In W.F. Bennett (Ed). Nutrient deficiencies and toxicities in crop plants. APS Press, St Paul, Minnesota.
Besalatpour, A.A., Ayoubi, S., Hajabbasi, M.A., Mosaddeghi, M.R. and Schulin, R. 2013. Estimating wet soil aggregate stability from easily available properties in a highly mountainous watershed. Catena, 111: 72-79.
Bocco, M., Willington, E. and Arias, M. 2010. Comparison of regression and neural networks models to estimate solar radiation. Chilean Journal of Agricultural Research, 70: 428-435.
Cecílio Filho, L.A.B., Feltrim, A.L., Mendoza Cortez, J.W., Gonsalves, M.V., Pavani, L.C. and Barbosa, J.C. 2015. Nitrogen and potassium application by fertigation at different watermelon planting densities. Journal of Soil Science and Plant Nutrition, 15 (4): 928-937.
Chrysargyris, A., Xylia, P., Botsaris, G. and Tzortzakis, N. 2017. Antioxidant and antibacterial activities, mineral and essential oil composition of spearmint (Mentha spicata L.) affected by the potassium levels. Industrial Crops and Products, 103: 202-212.
Chrysargyris, A., Xylia, P., Botsaris, G. and Tzortzakis N. 2017. Optimization of potassium fertilization/nutrition for growth, physiological development, essential oil composition and antioxidant activity of Lavandula angustifolia Mill. Journal of Soil Science and plant Analysis, 17(2).
Dai, F., Zhou, Q., Lv, Z., Wang, X. and Liu, G. 2014. Spatial prediction of soil organic matter content integrating artificial neural network and ordinary kriging in Tibetan Plateau. Ecological Indicators, 45: 184-194.
Dai, L., Li, Y., Wang, Q., Niu, J., Li, W., Wang, Y., Wang, J., Zhou, Z. and Gao, N. 2022. Chamomile: A Review of Its Traditional Uses, Chemical Constituents, Pharmacological Activities and Quality Control Studies. Molecules, 28(1). https://doi.org/10.3390/molecules28010133.
Emami, A.S. 1996. Methods of Plant Analysis, Volume II. Technical Journal No. 982. Karaj, Iran: Soil and Water Research Institute (In Persian).
Jat, R.S. and Ahaheat, I.P.S. 2006. Direct and residual effects of vermicomposts, biofertilizer and phosphorus on soil nutrient dynamics and productivity of chickpea-fodder maize sequence. Journal of Sustainable Agriculture, 28(1): 41-54.
Klute, A. 1986. Method of soil analysis, Part I, Chemical and microbiological.
Malakouti, M.J., MC Karimian, N. and Keshavarz, P. 2008. Determination Methods for Nutritional Deficiencies and Recommendations for Fertilizer. Office for the Publishing of Scientific Works. Tehran, Iran (In Persian).
Malakouti, M. C and Shahabi, A. 2000. Chloro-calcium spraying is an indispensable necessity in improving the qualitative characteristics of apple. Deputy Directorate for the Promotion of TAT. Ministry of Agriculture, Tehran. (In Persian).
Marashi, M., Mohammadi Torkashvand, A., Ahmadi, A. and Esfandyari, M. 2017. Estimation of soil aggregate stability indices using artificial neural network and multiple linear regression models. Spanish Journal of Soil Science, 7: 122-132.
Marashi, M., Mohammadi Torkashvand, A., Ahmadi, A. and Esfandyari, M. 2019. Adaptive neuro-fuzzy inference system: estimation of soil aggregates stability. Acta Ecologica Sinica, 39: 95-101.
Moghimi, S., Mahdian, M.H., Parvizi, Y. and Masihabadi, M.H. 2014. Estimating effects of terrain attributes on local soil organic carbon content in a semi-arid pastureland. Journal of Biodiversity and Environmental Sciences, 5(2): 67-106.
Mohammadi Torkashvand, A., Ahmadipour, A. and Khaneghah, A.M. 2020. Estimation of kiwifruit yield by leaf nutrients concentration and artificial neural network. The Journal of Agricultural Science, 158 (3): 185-193.
Mohammadi Torkashvand, A., Ahmadi, A. and Nikravesh, N. 2017. Prediction of kiwifruit firmness using fruit mineral nutrient concentration by artificial neural network (ANN) and multiple linear regressions (MLR). Journal of Integrative Agriculture, 16(7): 1634-1644.
Mohammadi Torkashvand, A., Ahmadi, A., Gómez, P.A. and Maghoumi, M. 2019. Using artificial neural network in determining postharvest life of kiwifruit. Journal of The Science of Food and Agriculture. DOI 10.1002/jsfa.9866.
Mokhtari Karchegani, P., Ayoubi, SH., Honarju, N. and Jalalian, A. 2011. Predicting soil organic matter by artificial neural network in landscape scale using remotely sensed data and topographic attributes. Geophysical Research Abstracts, 13, article no. EGU2011-1075. Available online from https://meetingorganizer.copernicus.org/EGU2011/EGU2011-1075.pdf (accessed 3 December 2019).
Niazian, M., Sadat Noori, S. and Abdipour, M. 2018. Artificial neural network and multiple regression analysis models to predict essential oil content of ajowan (Carum copticum L.).
Nurzynska-Wierdak, R. 2013. Does mineral fertilization modify essential oil content and chemical composition in medicinal plants? Acta Scientiarum Polonorum Hortorum Cultus, 12: 3-16.
Omid Beigi, R. 2004. Production and processing of medicinal plants, Volume 3. Astan Quds Razavi Press. 397 Page. (In Persian).
Pacheco, C., Calouro, F., Vieira, S., Santos, F., Neves, N., Curado, F., Franco, J., Rodrigues, S. and Antunes, D. 2008. Influence of nitrogen and potassium on yield, fruit quality and mineral composition of kiwifruit. International Journal of Energy and Environment, 2: 9-15.
Parvizi, Y., Gorji, M., Omid, M., Mahdian, M.H. and Amini, M. 2010. Determination of soil organic carbon variability of rainfed crop land in semi-arid region (neural network approach). Modern Applied Science, 4: 25-33.
Paye, A.L., Miller, R.H. and Keeney, D.R. 1948. Method of soil analysis, Part II., SSSA Inc.
Peng, J., Zhang, Y.Z., Pang, X.A. 2012. Hyperspectral features of soil organic matter content in South Xinjiang. Arid Land Geography, 55(5): 740–746.
Petronilho, S., Maraschin, M., Delgadillo, I., Coimbra, M.A. and Rocha, S.M. 2011. Sesquiterpenic composition of the inflorescences of Brazilian chamomile (Matricaria recutita L.): Impact of the agricultural practices. Industrial Crops and Products, 34: 1482–1490. doi: 10.1016/j.indcrop.2011.05.005.
Poorghadir, M., Mohammadi Torkashvand, A., Mirjalili, S.A. and Moradi, P. 2021. Interactions of amino acids (proline and phenylalanine) and biostimulants (salicylic acid and chitosan) on the growth and essential oil components of savory (Satureja hortensis L.). Biocatalysis and Agricultural Biotechnology, 30: 101815.
Radkowski, A. and Radkowska, I. 2018. Influence of foliar fertilization with amino acid preparations on morphological traits and seed yield of timothy. Plant, Soil and Environment, 64: 209-213.
Rahmani Khalili, M., Asadi, M.E., Mohammadi Torkashvand, A. and Pazira, E. 2020. Regression Analysis for Yield Comparison of Saffron as Affected by Physicochemical Properties of the Soil, Case Study in Northeast of Iran. Agricultural Research, 9(12). https://doi.org/10.1007/s40003-020-00455-6.
Savitikadi, P., Jogam, P., Khan Rohela, G., Ellendula, R., Sandhya, D., Rao Allini, V. and Abbagani, S. 2020. Direct regeneration and genetic fidelity analysis of regenerated plants of Andrographis echioides (L.) - An important medicinal plant, Industrial Crops and Products, 155, 112766.
Soltanpour, P.N. and Schwab, A.P. 1977. A new soil test for simultaneous extraction of macro- and micronutrients in alkaline soils. Communication in Soil Science and Plant Analysis, 8: 195-207.
Tashakori, F., Mohammadi Torkashvand, A., Ahmadi, A. and Esfandiari, M. 2020. Comparison of different metthods of estimating saffron yield based on soil properties in Golestan province. Communications in Soil Science and Plant Analysis, 51(13): 1767-1779.
Tofighi Alikhani, T., Tabatabaei, S.J., Mohammadi Torkashvand, A., Khalighi, A. and Talei, D. 2021. Effects of silica nanoparticles and calcium chelate on the morphological, physiological and biochemical characteristics of gerbera (Gerbera jamesonii L.) under hydroponic condition. Journal of Plant Nutrition, 44(7): 1039-1053.
Ubessi, C., Tedesco, S.B., de Bona da Silva, C., Baldoni, M., Krysczun, D.K., Heinzmann, B.M., Rosa, I.A. and Mori N.C. 2019. Antiproliferative potential and phenolic compounds of infusions and essential oil of chamomile cultivated with homeopathy. Journal of Ethnopharmacology, 239: 111907. doi: 10.1016/j.jep.2019.111907.
Upadhyay, R. K. and Patra, D.D. 2011. Influence of secondary plant nutrient (Ca and Mg) on growth and yield of chamomile (Matricaria recutita L.). Asian Journal of Crop Science, 3: 151-157.
Walkley, A.J. and Black, I.A. 1934. Estimation of soil organic carbon by the chromic acid titration method. Soil Science, 37: 29-38.
Wan, W.T., Song, Y.J., Xu, L.J., Xiao, P.G. and Miao, J.H. 2019. Research Review and Application Prospect Analysis of Matricaria. Modern Chinese Medicine, 21: 260–265. doi: 10.13313/j.issn.1673-4890.20181008002.
Zhou, T., Shi, P.J., Luo, J.Y. and Shao, Z.Y. 2008. Estimation of soil organic carbon based on remote sensing and process model. Frontiers of Forestry in China, 3: 139–147.
Zhao, D.S., Han, S.L., Yi, Y.J., Qiu, L., Ren, L.J. and Li, X.X. 2015. Standard operating procedure for standardized planting of local medicinal herb German chamomile.
doi: 10.13989/j.cnki.0517-6611.2015.28.028.Zhao, Y.F. 2018. Master’s Thesis. China Academy of Chinese Medical Sciences; Beijing, China: Study on Chemical Composition and Quality Standard of Uyghur Chamomile.