FULL-LENGTH RESEARCH ARTICLE Integrated Assessment and Modeling of Agricultural Mechanization in Potato Production of Iran by Artificial Neural Networks Morteza Zangeneh • Mahmoud Omid • Asadolah Akram Received: 4 June 2012 / Accepted: 24 January 2015 / Published online: 27 February 2015 Ó NAAS (National Academy of Agricultural Sciences) 2015 Abstract Mechanization is a concept and cannot be measured directly, so we used appropriate indicators in order to determine mechanization status for the first time in this field of study. In this paper, based on energy and power availability, four indicators, namely machinery energy ratio, mechanization index, productivity level of consumed power, and mechanization level, were selected for assessing agricultural mechanization in potato production of Iran using integrated assessment and modeling (IAM) to provide insight into the potential impacts of policy changes. To do an IAM in agricultural mechanization, we used pervasive analysis using more than 90 features in sample farms. This IAM is the first generalized model in agricultural mechanization. The main purpose of this study is presenting and showing capability of ANNs to model agricultural mechanization status and indicating best ANN model. Finally, a two hidden layer model with these features showed best performance: generalized feed-forward network with Levenberg Marquart learning rule and Bias Axon transfer function with 4–10 neurons in two hidden layers which have 27 input items for modeling four outputs. In this study, ANN models were introduced and applied to help IAM investigation which integrating production factors to have better knowledge about agricultural system of potato production in a wide region. Keywords Potato Á Integrated assessment and modeling Á Mechanization indicators Á Assessment Abbreviations ANN: Artificial neural networks; Ax: Axon (transfer function of ANN); BA: BiasAxon (transfer function of ANN); CG: ConjugateGradient (learning rule of ANN); DBD: DeltaBarDelta (learning rule of ANN); GFF: Generalized feed-forward; IAM: Integrated assessment and modeling; LA: LinearAxon (transfer function of ANN); LM: LevenbergMarquar (learning rule of ANN); LSA: LinearSigmoidAxon (transfer function of ANN); LTA: LinearTanhAxon (transfer function of ANN); MAE: Mean absolute error; MER: Machinery energy ratio; MI: Mechanization index; ML: Mechanization level; MLP: Multi layer perceptron; MoM: Momentum (learning rule of ANN); MSE: Mean squared Error; NMSE: Normalized mean squared error; PLCP: Productivity level of consumed power; QP: Quickprop (learning rule of ANN); R 2 : Coefficent of determination; SA: SigmoidAxon (transfer function of ANN); SMA: SoftMaxAxon (transfer function of ANN); TA: TanhAxon (transfer function of ANN) Introduction Agricultural Mechanization Better knowledge of the past and present is a key compo- nent for the improvement of the planning process that will impact Iran’s agricultural sector in the years to come. Findings from this and similar studies can be used to set M. Zangeneh (&) Á M. Omid Á A. Akram Department of Agricultural Machinery Engineering, Faculty of Agricultural Engineering and Technology, School of Agriculture & Natural Resources, University of Tehran, Karaj, Iran e-mail: mzangeneh@ut.ac.ir M. Omid e-mail: omid@ut.ac.ir A. Akram e-mail: aakram@ut.ac.ir 123 Agric Res (September 2015) 4(3):283–302 DOI 10.1007/s40003-015-0160-z