Modelling the Canopy Conductance of Cocoa Tree Using a Recurrent Neural Network
Opeyemi Samuel Sajo,
Philip Gbenro Oguntunde,
Johnson Toyin Fasinmirin,
Akindele Akinnagbe,
Ayorinde Akinlabi Olufayo,
Samuel Ohikhena Agele
Issue:
Volume 7, Issue 2, December 2021
Pages:
23-29
Received:
22 June 2021
Accepted:
10 July 2021
Published:
23 August 2021
Abstract: Direct measurement of crop water use is difficult and labour intensive. In some cases, the technicalities involved can only be exploited by well-trained researchers. Therefore, estimating this important crop parameter from readily available climatic data by way of modelling will ease the burden of direct measurement. The aim of the study is to parameterize models of canopy conductance of rain-fed cocoa tree, suitable for inclusion in physically-based model for predicting water use of cocoa trees. To do this, Sap flow density was monitored in three cocoa trees (Forestaro cultivar group) at the eight (8) year old cocoa plantation of the Federal University of Technology, Akure, Nigeria (7° 18' 15.9"N, 5° 07' 32.3"E), from 8th March 2018 to 7th March 2019, covering the two seasons of the region. Cocoa tree transpiration was determined from the measured sap flow and fitted into a physically based model (PM) to derive canopy conductance used for modelling. To choose the best model that predicts canopy conductance (the stomata control of water transport) in cocoa trees, Vector Autoregressive Models (VAR), a multivariate time series model, and Long Short-Term Memory (LSTM) network, an Artificial Intelligence (AI) model were employed. The prediction power of the VAR model was assessed and visualized using the vars R package, while the LSTM model, a Recurrent Neural Network (RNN) algorithm was implemented using Python programming within Google COLAB jupyter notebook. Before modelling, data were tested for stationarity using the Augmented Dickey-Fuller test. While two-thirds of the data were used to train the models, the remaining one-third of the data were used to test the trained model. As VAR models were evaluated using R-squared and Root Mean Squared Error (RMSE), LSTM was evaluated by comparing the train loss and test loss, and also RMSE. VAR (with Adjusted R-Squared=0.11) is found not to be suitable to model the complex relationship between canopy conductance and climatic variables. Further iteration to exclude insignificant climatic variables from the VAR model did not also improve the model. However, LSTM with RMSE of 0.026 and having the test loss not dropping below the training loss was observed to perform better in modelling the canopy conductance of Cocoa. The result of the research further revealed that temporal dynamics of transpiration is complex and difficult to be defined by traditional regression. LSTM with a prediction accuracy of 97.4% could therefore be used for the prediction of cocoa canopy conductance.
Abstract: Direct measurement of crop water use is difficult and labour intensive. In some cases, the technicalities involved can only be exploited by well-trained researchers. Therefore, estimating this important crop parameter from readily available climatic data by way of modelling will ease the burden of direct measurement. The aim of the study is to para...
Show More
Forecasting Foodgrains Production Using Arima Model and Neural Network
Veluchamy Kasthuri,
Subbiah Selvakumar
Issue:
Volume 7, Issue 2, December 2021
Pages:
30-37
Received:
17 January 2021
Accepted:
30 January 2021
Published:
31 August 2021
Abstract: The time series is a set of values arranged in a specific order of time. Prediction and analysis of food grain is a vital role in agriculture statistics. The Agriculture Statistics System is very complete and provides data on a wide range of topics such as crop area and production, land use, irrigation, land holdings, agricultural prices and market intelligence, livestock, fisheries, forestry etc. Agricultural credit and subsidies also consider important supporting factors for agricultural growth. India is the world's largest producer of millets and second-largest producer of wheat, rice, and pulses. The present research work focused on production of food grains in India using time series data ranging from 1990- 91 to 2018-19. In this paper, Autoregressive Integrated Moving Average Model (ARIMA), Multilayer Perceptron (MLP) and Radial Basis Function (RBF) for predicting foodgrains of India were compared. Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) were compared. The results were displayed numerically and graphically.
Abstract: The time series is a set of values arranged in a specific order of time. Prediction and analysis of food grain is a vital role in agriculture statistics. The Agriculture Statistics System is very complete and provides data on a wide range of topics such as crop area and production, land use, irrigation, land holdings, agricultural prices and market...
Show More
Communication Between Neural Networks, and Beginning of Language
Issue:
Volume 7, Issue 2, December 2021
Pages:
38-44
Received:
8 December 2021
Accepted:
21 December 2021
Published:
31 December 2021
Abstract: There is a view that the cranial nerve circuit is composed of a combination of the same modules as the basic functions. According to that view, the author has presented the module (Basic Unit) that performs parallel-serial mutual conversion and has shown that the neural network that recognizes and generates arbitrary time-series data can be constructed by combining the module. In Chapter 2, the neural network that has the functions of federated learning and imitation that enable collective behavior of animals is shown, and added an idea of concrete circuit configuration to published papers. In Chapter 3, following a consideration of the fundamental role of language, a neural network with the same basic structure connected to the upper level of the neural network shown in Chapter 2 but with functions closely related to language is presented. The new neural network consists of a pair of neural networks that handle languages and images respectively. Each activated part is expressed using the Category theory concept. Category's entity is a set of Basic Units connected each other and changes of their state. The activated Categories are tied with the corresponding activation part in the pairing neural network, and interconverting is performed. The state of the Basic Unit may be inspired by sensory organs, but behave independently of the actuating behavior of conventional neural networks connected to the low position. Humans can generate an image of events that may occur in past or in future even if that are not directly related to the situation in front of the eye, and share their images by dialogue. The dialogue consists of time series data with a response format such as question or negation. The newly added neural network helps generate shared recognition.
Abstract: There is a view that the cranial nerve circuit is composed of a combination of the same modules as the basic functions. According to that view, the author has presented the module (Basic Unit) that performs parallel-serial mutual conversion and has shown that the neural network that recognizes and generates arbitrary time-series data can be constru...
Show More
