Modelling of soil temperature by ann

In this work, Artificial Neural Network (ANN), which is a computational model based on the structure and functions of biological neural networks, is used to solve the problems being faced by the farmers. The main aim of the thesis is to increase farm productions. Some parameters hindering the productivity of the farmers were checked and such parameters were highly and carefully looked into in this research work. We implemented two methods in this research in order to investigate and classify the low farm productivity hindering parameters such as, soil quality, soil pH-Value, soil water retention capacity, weather and atmospheric conditions etc. Two methods; "Curve fitting" and "Pattern classification", we're implemented to investigate the cause of low farm productivity. Literally, we can say that "Curve-fitting" is the process of constructing a curve, or mathematical function, that has the best fit to a series of data point, possibly subjected to constraints. It can involve either interpolation, where an exact fit to the data is required, or smoothing, in which a smooth function is constructed that approximately fits the data. "Curve fitting method" also uses the approach of representing an attempt for the neural network to identify and approximate an arbitrary input-output relation and once the relation has been modelled, the necessary accuracy by the network can be used for a variety of tasks. We use Artificial Neural Network (ANN) to investigate and suggest possible solutions for solving problems. Generally, two novel methodological approaches will be applied on raw data in this study. We believe that the increasing ratio in the productivity of farm produce can better be enhanced. If careful study to significantly sort out, these problems are urgently looked into. We will lead to more farmers' productivity. In this research, we used collected data form a single station located in Istanbul, (Akören) Turkey, at geographical coordinates are 41,1568 oN - 28,3119 oE, dating from 01.01.2016 to 31.12.2016. Data is recorded every ten (10) minutes of intervals. To define the role of soil temperature on farm production; we consider Surface Temperature (5 cm), Solar Radiation, UV Radiation, Wind Speed and Direcction. These parameters are based on TARSİM'e (Tarım Sigortaları) archive data. The first method we employed is "Curve-fitting method", which is used in modeling the necessary data and parameter accuracies by the network. It can further be used for a variety of tasks, such as series prediction, function approximation, and function optimization. "Curve-fitting method" has the objective of selecting parameter values which minimizes the total error over the set of data points being considered. We investigated that the second method could help for some possible solution to the problems faced by the farmers is "Pattern classification". It is a scientific discipline aimed at classifying objects into a set of categories or classes. It is also regarded as a sub-topic of machine learning concerned with the automatic discovery of regularities in data through the use of learning algorithms. The learning algorithms involves building a function that can map the input feature space to an output space of two or more classes. The goal of pattern classification is to assign input patterns to one of a finite number of classes. This method has been tested in previous research and gave positive results. The simulation results gotten depicts that "Levenberg-Marquardt (LM)" and scaled conjugate gradient (SCG) methods are the best of the seven (7) previous methods that has been used so far in this field of research by other researchers. The accuracy of the networks studied, ranged from 94.4 to 99.2% and we observed that networks trained with LM algorithm were faster. Furthermore, these models are sophisticated and we realized that they can predict the status of the soil and also inform the farmers with the right decision to protect their fields.