Artificial intelligent system for diagnosis and management of maize pest in Uasin Gishu county, Kenya
Nyang’anga, Thadias H
MetadataShow full item record
Agriculture contributes 24.5percent of the Kenyan Gross Domestic Product (GDP). This makes it one of the economic pillars of the country‘s Economy (Government of Kenya, Agricultural Sector & Strategy, 2010). Maize is the largest consumed staple food in Kenya. However, losses are always incurred by farmers due to diseases and pest infestation throughout the plants life cycle. It was evident that proper control and management of pest and diseases affecting crops will increase production output and hence economic growth. Pest diagnostics and management services are currently offered by the Government to farmers through the extension officers. However, delivery of these services is hindered by several factors such as understaffing, poor infrastructure and reduced budgetary allocation to agriculture by the Government of Kenya at a time when the operation costs are on the increase. Due to these challenges, extension officers are not in a position to deliver diagnostics and disease management services to the farmers efficiently, timely and effectively. On the other hand, there has been an increased investment by the Government on Information and Communication Technology (ICT) infrastructure and increase usage of mobile telephones by the Kenyan population, it was important to consider the use of information communication technologies to deliver diagnostics and management services to farmers efficiently, timely and cost effectively. The study was carried out to investigate how artificial intelligent systems could be used for agricultural diagnostics and management of Maize diseases in Uasin Gishu County, Kenya. The research design was a case study of maize pests, extension officers and maize farmers in Uasin Gishu County. The population targets for the study were maize farmers, extension officers, and publications on maize pests. The sample size of the farmers was determined using the formula by Mugenda and Mugenda, (2003). No sampling procedure was used on extension officers and publications on Maize pest because of their low numbers hence data was collected from the whole population. Sample size of ninety farmers, eighteen extension officers and twelve publications was used. Interview guides was used to collect primary data from the farmers and extension officers while forms were used to collect secondary data from publications. The data was statistically analysed using SPSS software and results presented. In the case of developing an electronic database on existing knowledge and practices on maize disease diagnostics and management, secondary data was collected on the various maize diseases and their management from published work. An open electronic database was developed using mySQL and the resulting database schema, entity relationship diagram, relational tables and database architecture presented. The artificial intelligent system was developed and implemented using the Object oriented analysis and design methodology. The results showed that both farmers and extension officers wanted a system which is web based and could be accessed using mobile phones and computers through the mobile interface and main site respectively. They also required the query interface to have both pictures and textual description of symptoms for ease of the selection options. Printable format and full diagnostic information content was preferred by both the farmers and extension officers. Full information on the various maize disease which include name, symptoms, picture of every symptom, available management options, chemical/drugs to administer, how to administer and where to purchase the chemicals from, were found and used in the building of the open database. The modules from MICDES of retrieve, reuse, revise and retrain and algorithms were found to be the appropriate module for the artificial intelligent system. The system was developed, tested and hosted. The system was found to provide accurate diagnostics service. It was therefore concluded that artificial intelligent system can be used to provide extension services such as diagnosis and disease management information sharing.