ISHS Acta Horticulturae 230: Symposium on High Technology in Protected Cultivation
A KNOWLEDGEBASE SYSTEM AND A NEURO-COMPUTER SYSTEM FOR AGRICULTURAL INFORMATION PROCESSING |
| Authors: | M. Hirafuji, Y. Ono, K. Kobayashi |
| Abstract:
We developed an original knowledge-based system (Protaios) and a Neuro-Computing System(NCS). They are for control, management and decision support in agricultural production. Protaios is a kind of conventional expert system shell, and NCS is a "neuro-computing shell" in a word. Protaios has the function of understanding simple natural English based on the technique of natural language processing. With this function, rules for knowledgebases can be generated automatically from English documents in literatures and databases. This function serves for building huge knowledgebase and a kind of common knowledge bank over the wide domain. Some knowledgebases were developed easily with Protaios. But the conventional AI methods can not avoid the difficulty to develop advanced control system such as self controlled agri-robots and automatic plant factories because of slow execution time and difficulty of adjusting the weights of uncertainty in rules. We think the power of neuro-computing solves all these problems, so we developed an original neuro-computing system on a personal computer. The artificial neural network have many functions such as learning, pattern recognition, association and analogy. We have applied N C S to the some problems such as pattern recognition of image and time series data, nonlinear multivariate analysis, "neural expert system" and so on, which are difficult for conventional methods. And we confirmed the ability of the both tools to many problems in agricultural information processing through actual demonstrations. | |
Download Adobe Acrobat Reader (free software to read PDF files) | |
