Avaliação de conhecimento estruturado no sistema de apoio a decisão para semiologia do cafeeiro
Data
2011-08-05
Autores
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Federal de Lavras
Resumo
O Brasil é o maior produtor mundial de café e o segundo maior mercado consumidor, portanto, a cultura do cafeeiro tem grande importância cultural e socioeconómico para o país. Entre os limitantes na produtividade estão, principalmente, as doenças, as pragas, as deficiências nutricionais e fatores abióticos. O diagnóstico preciso e confiável é fundamental para o controle das doenças, além de evitar contaminações ambientais com a utilização desnecessária de defensivos agrícolas. Muitas pesquisas são realizadas todos os anos gerando novos conhecimentos e técnicas que contribuem para o aumento da produtividade. Todo este conhecimento e técnicas geradas precisam ser difundidos e a informática oferece algumas soluções, entre elas o emprego de sistemas de apoio à decisão (SAD), que atua disponibilizando o conhecimento e reduzindo possíveis erros de diagnose e manejo. Para disponibilizar o conhecimento acumulado nos últimos dez anos e divulgá-lo de forma eficiente, este trabalho foi realizado com os seguintes objetivos: i) construir uma base de conhecimentos para diagnóstico de doenças bióticas e abióticas; ii) desenvolver e implementar a nova versão do SAD e iii) avaliar o sistema. O conhecimento foi adquirido por meio de entrevistas, com especialistas e, para complementar, foram consultadas literaturas específicas sobre a cafeicultura. No desenvolvimento da nova versão do SAD foi construída a interface em ambiente atual (Windows Seven), inteiramente ilustrada. O SAD foi construído utilizando-se a ferramenta ‘Borland Delphi’. Foram inseridos 35 novos diagnósticos (3 de doenças, 5 de pragas, 6 de causas nutricionais e 21 de causas abióticas), 34 novos ramos e 66 nós. Foi dada ênfase nas fotografias, inserindo- se 20 novas fotografias de sintomas e 36 ilustrações. A validação do programa foi realizada por três grupos, de diferentes níveis de conhecimento (10 pessoas/grupo), tentando diagnosticar corretamente. Obteve-se um acerto de 23% dos usuários e 91% do SAD. O sistema foi denominado Doctor Coffee v2.0.
Brazil is the largest producer world's of coffee and the second largest consumer market, so the culture of coffee has great socioeconomic and cultural importance for the country. Among the limiting factors in productivity are mainly diseases, pests, nutritional deficiencies and abiotic factors. The reliable and accurate diagnosis is critical to disease control, and prevent environmental contamination with unnecessary use of pesticides. Many surveys are conducted every year generating new knowledge and techniques that contribute to increased productivity. All this knowledge and techniques need to be generated and distributed so computing offers some solutions, including the use of decision support systems (DSS), which acts providing knowledge and reducing possible errors in diagnosis and management. To make the knowledge accumulated in the last ten years, and disseminate it efficiently, this work was performed with the following objectives: i) building a knowledge base for diagnosis of biotic and abiotic ii) develop and implement the new version of DSS and iii) evaluate the system. The knowledge was gained through interviews with experts and for complementary literatures were consulted on the specific coffee. During the development of the new version of the DSS the interface was built in the current environment (Windows Seven) and fully illustrated. The DSS was built using the tool “Borland Delphi. ". 35 new diagnosis were included (3 diseases, 5 pests, 6 nutritional causes and 21 of abiotic causes), 34 new branches and 66 knots. Emphasis was placed on the photographs, added to 20 new photographs and 36 illustrations of symptoms. The validation program was conducted by three groups, different levels of knowledge (10 peaple / group), trying to diagnose correctly. We obtained an accuracy of 23% and 91% of users of the DSS. The system was called “Doctor Coffee v2.0”
Brazil is the largest producer world's of coffee and the second largest consumer market, so the culture of coffee has great socioeconomic and cultural importance for the country. Among the limiting factors in productivity are mainly diseases, pests, nutritional deficiencies and abiotic factors. The reliable and accurate diagnosis is critical to disease control, and prevent environmental contamination with unnecessary use of pesticides. Many surveys are conducted every year generating new knowledge and techniques that contribute to increased productivity. All this knowledge and techniques need to be generated and distributed so computing offers some solutions, including the use of decision support systems (DSS), which acts providing knowledge and reducing possible errors in diagnosis and management. To make the knowledge accumulated in the last ten years, and disseminate it efficiently, this work was performed with the following objectives: i) building a knowledge base for diagnosis of biotic and abiotic ii) develop and implement the new version of DSS and iii) evaluate the system. The knowledge was gained through interviews with experts and for complementary literatures were consulted on the specific coffee. During the development of the new version of the DSS the interface was built in the current environment (Windows Seven) and fully illustrated. The DSS was built using the tool “Borland Delphi. ". 35 new diagnosis were included (3 diseases, 5 pests, 6 nutritional causes and 21 of abiotic causes), 34 new branches and 66 knots. Emphasis was placed on the photographs, added to 20 new photographs and 36 illustrations of symptoms. The validation program was conducted by three groups, different levels of knowledge (10 peaple / group), trying to diagnose correctly. We obtained an accuracy of 23% and 91% of users of the DSS. The system was called “Doctor Coffee v2.0”
Descrição
Dissertação de Mestrado defendida na Universidade Federal de Lavras
Palavras-chave
Diagnose, Manejo
Citação
SANTOS, L. A. Avaliação de conhecimento estruturado no sistema de apoio a decisão para semiologia do cafeeiro. 2011. 56 f. Dissertação (Mestrado em Agronomia/Fitopatologia) - Universidade Federal de Lavras, Lavras. 2011.