Predicting aflatoxin levels a spatial autoregressive approach
- 1. Predicting Aflatoxin levels: An Spatial Autoregressive approach Gissele Gajate-Garrido, IFPRI International Food Policy Research Institute Uniformed Services University of the Health Sciences International Center for the Improvement of Maize ACDI/VOCA/Kenya Maize Development Program and Wheat Kenya Agricultural Research Institute International Crops Research Institute for the Semi- Institut d’Economie Rurale Arid Tropics The Eastern Africa Grain Council University of Pittsburgh
- 2. Collecting aflatoxin information is time consuming and expensive. Sometimes we can have aflatoxin information from a smaller sample of households. These information could be useful to predict the level of aflatoxins in other households with similar characteristics.
- 3. A Spatial Autoregressive Model (SAR) uses the household characteristics and the aflatoxin level of people around it to predict aflatoxin levels in each household.
- 4. This model gives more weight Aflatoxin level to the information of my closest “neighbors” and less to the ones that are further away. My “neighbors” information could help predict my own Observable Unobservable: aflatoxin level since it could characteristics - Attitudes - Risk aversion contain information that - Motivation usually is not captured by surveys. When we estimate models there is always an error term present that represents the variation that we are unable to capture.
- 5. There are variables such as a person’s determination or innate ability that could help predict how much time and effort they will invest in preventing aflatoxins in their crops. These variables cannot be observed or recorded in a survey. However, by capturing information about my peers this could help provide additional information about how I behave and how high is my aflatoxin level.
- 6. In order to asses who is “closest” to me I use location variables: Longitude Latitude Elevation Slope ▪ (Only for the pre-harvest sample)
- 7. 90% Storage 80% 70% 63% 74% 60% Production 50% 38% 40% 29% 27% 30% 20% 6% 9% 9% 10% 6% 2% 0% Treated Improved Pesticide Fertilizer Insect Rodent Plastic Storage: Frequent Hand soil (lime, seed damage damage bags for special use of sorting manure, storage room pestcide before etc.) inside in storage house storage
- 8. We use 100% Aflatoxin variation data from 90% 80% 36% Mali to 70% The inside sample prediction captures test the 60% 36% of the variation in prevalence model. 50% values. 40% 64 % Yet, the information of my neighbors is not useful to predict my prevalence We start 30% levels, only my characteristics are with pre- 20% relevant. 10% harvest 0% data. My neighbors' My characteristics Unobservable
- 9. 2.5 2 The relationship 1.5 between 1 1.04 *** predicted and 0.5 real values is 0 almost 1 to 1. 0 1 2 3 It is significant at Measured prevalence (part per billion) 1%. Predicted prevalence 45 degree line Variable Obs Mean Std. Dev. Min Max Measured prevalence 247 27.2 64.0 0.05 492.0 Predicted prevalence 247 29.6 26.9 0.00 130.7
- 10. Kernel density estimate for Pre-harvest Aflatoxin levels .04 76% The model is not .03 able to capture Density extremely high .02 values of prevalence and in general 43% .01 overestimates lower values. 0 0 20 100 200 300 400 500 prevalence (part per billion) Kernel density measured prevalence Kernel density predicted prevalence kernel = epanechnikov, bandwidth = 3.8288
- 11. Kernel density estimate for Main HH Pre-harvest Aflatoxin levels .01 .008 .006 Density 37% .004 63% .002 0 0 20 50 100 150 200 250 Kernel density predicted prevalence for Main HH kernel = epanechnikov, bandwidth = 12.9933 Variable Obs Mean Std. Dev. Min Max Predicted prevalence for main HH survey 1169 58.4 59.3 0.0 223.1
- 12. Post-harvest data after 1 month in Total variation in aflatoxin levels storage During storage Variation Variation not only your explained by explained by characteristics but personal neighbors characteristics aflatoxin level also your "neighbors" information help Unexplained variation = 62 % explain your aflatoxin level. The inside sample prediction captures 38% of the variation in prevalence values.
- 13. 2.5 2 The 1.5 relationship between 1 0.95 *** predicted and 0.5 real values is almost 1 to 1. 0 0 1 2 3 It is significant Measured prevalence (part per billion) at 1%. Predicted prevalence 45 degree line Variable Obs Mean Std. Dev. Min Max Measured prevalence 243 121.9 256.9 0.0 1911.2 Predicted prevalence 243 129.0 130.5 0.0 778.0
- 14. The same methodology applied to the data in Mali will be applied to the data in Kenya. Hence will be able to predict prevalence levels for the main household survey and use it for further analysis. Should we expect similar results? Different crops ▪ Mali –groundnuts vs. Kenya – maize It also depends on production and storage practices in Kenya.
- 16. We have two models that can be used to predict aflatoxin models: Maxent SAR model We need to compare the strengths and weakness of both models. We can also consider introducing other variables to improve the predictions.
- 17. Current Partners: Donor: Bill and Melinda Gates Foundation Center/ Universities IFPRI: C. Narrod (Project lead), P. Trench(Project manager), M. Tiongco, D. Roy, A. Saak, R. Scott, W. Collier, M. Elias. CIMMYT: J. Hellin, H. DeGroote, G. Mahuku, S. Kimenju, B. Munyua ICRISAT: F. Waliyar, J. Ndjeunga, A. Diallo, M. Diallo, V. Reddy University of Pittsburgh: F. Wu, Y. Liu US Uniformed Health Services: J. Chamberlin, P. Masuoka, J. Grieco Country Partners ACDI/VOCA: S. Collins, S. Guantai, S. Walker Kenya Agricultural Research Institute: S. Nzioki, C. Bett Institut d’Economie Rurale: B. Diarra, O. Kodio, L. Diakite