Agricultural Information Research
Online ISSN : 1881-5219
Print ISSN : 0916-9482
ISSN-L : 0916-9482
Original Paper
A Java Framework for Developing a Plant Growth and Disease Prediction Model
Kei Tanaka
Author information
JOURNAL FREE ACCESS

2006 Volume 15 Issue 2 Pages 183-193

Details
Abstract

In developing plant growth models, disease and pest forecasting models, the cost for developing acquisition functions of weather data, and various GUI components such as parameter setup and result display is higher than for developing calculation part which is main part of a model. If a model developer can use framework for model development, it is clear that development efficiency is improved by reducing programming for the common function of a model and concentrating on development of the calculation part of a model, from experience which developed several models in Java.
The framework provides the following functions; the function for treating date and time series data as model data, the function for using MetBroker which can access various databases by consistent methods and processing a user's data file, the GUI components for displaying result data with a table, a chart, and a map. These functions can be used only by description of several program lines which registers the key of data into model data as a data element.
For example, in development with the framework of MetBLASTAM which is a rice blast forecasting model, the 960 lines (except comment lines) of program was newly developed for the model. The total of the model program and the program included in the framework which called by the model are 17361 lines. The program newly developed is 5% of the whole program. This shows the efficiency of model development was improved and reusability of program is high.

Content from these authors
© 2006 Japanese Society of Agricultural Informatics
Previous article Next article
feedback
Top