Agricultural Data Mining the Basics - Identifying Farmers Profiles who Have the Greatest Risk for Severe Striga Infestations by`Baseline` Analysis: A Trans-disciplinary Overview of the Farm System

Dr. Roger R. MacLean
To add a paper, Login.

The data collection methodology, approach and statistical design is based on Weber`s theory of Abstraction which requires the identification and collection of the broadest number of variables in order to develop Ideal Farm Types. The first step was to develop an understanding of the farm system by collecting all forms of data; second step was to juxtapose and blend together the various forms of data in linear forms against a test variable Striga infestation levels; third step was to evaluate if the amount of knowledge gained in predicting Striga infestation levels was statistically significant by cross correlating soil nutrient levels, crop management approaches, farmers` perceptions of Striga infestation and spatial distances; fourth was to use parametric and non-parametric analytical tools in conjunction with data compression to locate the best combination of parameters to better manage Striga. The final part was to identify and integrate the crop, field and social data into farmer’s who have the highest and lowest likelihood of being infested by Striga. Results indicate that social and natural science data can be successfully cross correlated and combined if collected correctly. Found a correlation between soil nutrients and farmers` risk tolerance.

Keywords: Agriculture, Farm System, Mali West Africe, Striga, Sociology, Plant Science, Soil Science, Statistics, Farm Management
Stream: Sociology and Geography, Natural, Environmental and Health Sciences, Research Methodologies, Quantitative and Qualitative Methods
Presentation Type: Virtual Presentation in English
Paper: A paper has not yet been submitted.

Dr. Roger R. MacLean

Professor, Department of Sociology & Anthropology, CEGEP John Abbott College/ Concordia University
Montreal, Quebec, Canada

Full farm system analysis; Agriculture, Ecosystem, Agro-ecosystem integration and Sociology. Developed a methodological approach (modified Rapid Rural Appraisal) which allowed for the simultaneous collection and integration of natural science data and sociological (social science data) of African farms in terms of weed management (Striga). The natural science data included soil, plant science, climate data, weed data, crop management approaches and other aspects such as mineral composition. The social science data included family, status, economic, technological usage, farmer risk assessment, gender issues and agricultural usage systems. Developed statistical approaches which combined and integrated Sociological approaches (PRE Proportional reduction of error) such as cross tabulation, recoding or data compression, factor analysis combined with linear approaches. Findings found better ways to manage Striga and indications. Present interests will be to use Segmentation Modelling to analyse the farm system in terms of the impact of HIV.

Ref: I06P0348