Discriminant Analysis for Classifying Farm Size Based on Farm Management Variables
Article Sidebar
Main Article Content
Abstract
The aim of this study was to survey the background information of 120 dairy cattle farmers in the South-Central area. The information of individual farm managements was collected by interview schedule during 4 to 17 March, 2011. Discriminant analysis with proportional priors was used to analyze the effect of farm managements on the size of farms including small, medium and large farm sizes depend on the number of milking cows (n) within farms including n<15 (8 farms), 15≤n<30 (54 farms) and n≥30 (58 farms), respectively. Eight variables based on farm managements were analyzed and found that 7 variables including sex of farmer, age of farmer, number of family members, time duration to be the member of cooperation, blood level of Holstein Friesian, the amount of concentrate feed and the amount of roughage were significantly different (P<0.20) while percentage of concentrate feed was not significantly different (P = 0.225). The discriminant analysis with proportional priors was used to classify farms into the group of 3 different farm sized based on 8 variables. The result showed that the predictive discriminant model has ability to classify farms in to the corrected group (P<0.20) with Wilks'Lambda = 0.697, chi-qquare = 37.378, df = 20, P = 0.011 and canonical correlation = 0.465. The highest values of standardized canonical discriminant function coefficient were blood level of Holstein Friesian (0.836), the amount of concentrate feed (0.622), percentage of concentrate feed (0.462), number of family members (0.450), age of farmer (0.304), time duration to be the member of cooperation (0.279), sex of farmer (0.158) and the amount of roughage (0.158), respectively. The misclassification of the predictive discriminant model was 39.6%. The result showed that the most of cattles in the farms were less than or equal to 87.5% of Holstein blood. It might be that cattles with less than 87.5% of Holste Friesian blood suitable to survive in the environment in the South-Central Area.
Article Details
References
วลัยกานต์ เจียมเจตจรูญ และ วรรณา อ่างทอง. 2541. รายงานเบื้องต้นสภาพการจัดการด้านอาหาร โคเนื้อและโคนมของเกษตรกรในพื้นที่ 11 จังหวัดภาคกลาง. รายงานผลงานวิจัยประจำปี 2541 กองอาหารสัตว์ กรมปศุสัตว์ กระทรวงเกษตรและสหกรณ์. หน้า 295-313.
สมศักดิ์ เภาทอง ประเทศ ปุ้ยพันธวงศ์ และ วิโรจน์ ฤทธิ์ฤาชัย. 2541. การสำรวจสภาพการเลี้ยงโคนมของเกษตรกร ตำบลพังขว้าง อำเภอเมือง จังหวัดสกลนคร. รายงานผลงานวิจัยประจำปี 2541 กองอาหารสัตว์ กรมปศุสัตว์ กระทรวงเกษตรและสหกรณ์. หน้า 12-32.
Hossain, M., S. Wright and L. A. Petersen. 2002. Comparing performance of multinomial logistic regression and discriminant analysis for monitoring access to care for acute myocardial infarction. Journal of Clinical Epidemiology 55: 400-406.
Millogo, V., G. A. Ouedraogo, S. Agenas and K. Svennersten-Sjaunja. 2008. Survey on dairy cattle milk production and milk quality problems in peri-urban areas in Burkina Faso. African Journal of Agricultural Research 3(3): 215-224.
Montgomery, M. E., M. E. White and S. W. Martin. 1987. A comparison of discriminant analysis and logistic regression for the prediction of coliform mastitis in dairy cows. Canadian Journal of Veterinary Research 51: 495-498.
Mureda, E. and Z. M. Zeleke. 2008. Characteristics and constraints of crossbred dairy cattle production in lowland areas of Eastern Ethiopia. Livestock Research for Rural Development. 20(4): article 57 http://www.lrrd.org/lrrd20/4/mure20057.htm (December 20, 2011).
Riveiro-Valino, J. A., C. J. Alvarez-Lopez and M. Fco. Marey-Perez. 2009. The use of discriminant analysis to validate a methodology for classifying farms based on a combinatorial algorithm. Computers and Electronics in Agriculture 66: 113-120.
Tadesse, M. and T. Dessie. 2003. Milk production performance of zebu, Holstein Friesian and their crosses in Ethiopia. Livestock Research for Rural Delvelopment 3(15).
