Identification of Tomato Diseases using Skip-gram and LSTM Based on QA(Question-Answer) System

Abstract

In the field of agricultural information processing, automatic identification and diagnosis of common diseases of tomatoes play an important role. Deep learning is a hot research topic in the field of pattern recognition and machine learning. It can effectively solve some problems of vegetable pathology, such as disease identification, automatic control and production prediction. In this paper, the tomato pest and disease experiments were carried out with natural language data set of 1.12G pathological and healthy tomatoes crawled from Doctors of Agriculture Website as source data. Five common tomato diseases' symptoms were identified by training Skip-gram algorithm. Finally, symptoms identified above can be classified using LSTM algorithm with classifiers. The research shows that the tomato pest and disease corpus identification models based on LSTM algorithm with classifiers achieve an average accuracy which are over 60%. The simulation results of tomato diseases identification show feasibility and effectiveness of the methods. This article aims at integrating our methodology into working systems that can be used in the identification fields.