Electronic feeders are used to automatically measure individual feed intake on group-housed pigs and are often used in breeding herds. Data from electronic feeders contain errors that must be edited to improve data accuracy. The objectives of this research were to develop criteria to identify errors in data from electronic feeders, to identify factors that contribute to errors, and to develop and evaluate editing methods. An additional objective was to compare performance of growing pigs on electronic versus commercial multi-space feeders. Data from 591 pigs were used to develop 16 criteria to detect errors in visits. Percentage of visits that contained errors ranged from 4.3 to 18.7% in three data sets. Logistic regression analysis was used to identify factors that contribute to the frequency of each error type. Although errors were affected by replicate, sex, test day, feeder, and pig, the largest effect was week to week variation within feeder, which is affected by feeder management. A linear mixed model with 31 covariates related to error types was fit to daily feed intake without error visits and used to adjust daily feed intake for errors. This and five other editing methods were then used to edit data from 124 pigs with simulated errors. Accuracy of each editing method was evaluated by correlating edited feed intake from the data set with simulated errors to true feed intake. Accuracy ranged from 0.93 to 0.99 and from 0.82 to 0.96 for average daily feed intake and daily feed intake. The new editing method had the highest correlations for both traits and was robust in short and long test period and in data sets with varying amounts of errors. To evaluate the effect of electronic feeders on performance, Yorkshire boars and gilts (n = 475) were fed using a single-space electronic feeder or a five-space conventional feeder. Body weight, backfat thickness, loin muscle area, and pen daily feed intake per pig were measured serially. Gilts on electronic feeders ate less, grew slower, and deposited less backfat and loin muscle area throughout the test period. There were no differences for boars. In conclusion, with careful editing, accurate data can be obtained from electronic feeders. These data are useful for genetic selection related to feed intake, although caution should be used when testing gilts on electronic feeders because performance may be hindered.
The use of data from the National Pork Board for this research is gratefully acknowledged.
Keywords: Electronic Feeders, Feed Intake, Editing Methods, Performance