3.2. Comparison video feeding durat

3.2. Comparison video feeding duration and RFID registrations
In total 14 070 RFID registrations of the focal pigs occurred during
the observation time, with 704 ± 363 registrations per pig
(mean ± SD) (Table 1) on average. The scatter plot of the number
of registrations per pig against the feeding duration as captured
on video is illustrated in Fig. 3. If all pigs are considered, a coefficient
of determination R2 of 0.53 is obtained. This low fit is mainly
due to four pigs where the residual after linear regression is outside
the 95% confidence interval (i.e., pigs 11, 14, 19 and 20, hereafter
called ‘outliers’) – without those four pigs the R2 increases to
0.88. In both cases, overall fit was significant (P < 0.001).
To illustrate the correspondence between the video based
observations and RFID registrations for different pigs and to find
an explanation for the outlier pigs, the signals registered for a
14 min period containing feeding are illustrated in Fig. 4 for pigs
11, 13 and 14 (2 outlier pigs and a pig on the regression line). It
can be clearly seen that the RFID registrations during feeding for
pig 11 are systematically lower than the video registrations, while
those for pig 13 agree and those for pig 14 are higher. This explains
why in Fig. 3, pig 11 is below the regression line, pig 13 is on the
line and pig 14 is above it. The lower RFID registrations for pig
11 and 19 (outliers on Fig. 3) can be explained by these pigs’ unusual
habit of often lying down while feeding. These two pigs were
the only focal pigs displaying this behaviour multiple times during
the video observations. This prone position increased the distance
between the RFID tags in their ears and the RFID antenna above
their heads, which therefore may have reduced the number of registrations
All four of the outliers (pigs 11, 14, 19 and 20) showed different
feeding behaviour than the rest of the focal pigs during the observation
period. This explains their outlying value in the scatter plot
(Fig. 3). By examining the number of RFID registrations per day
during the entire fattening period for pigs 11, 14, 19 and 20 compared
to those of pig 13, we can deduce whether this feeding
behaviour is consistent or not (data not shown). The number of
registrations of pig 11 and pig 19 were typically lower than those
of pig 13, while those of pig 14 and 20 were typically higher than
those of pig 13. This was especially the case in the 30 days around
the observation day.

ผลลัพธ์ (ไทย) 1:

3.2. Comparison video feeding duration and RFID registrationsIn total 14 070 RFID registrations of the focal pigs occurred duringthe observation time, with 704 ± 363 registrations per pig(mean ± SD) (Table 1) on average. The scatter plot of the numberof registrations per pig against the feeding duration as capturedon video is illustrated in Fig. 3. If all pigs are considered, a coefficientof determination R2 of 0.53 is obtained. This low fit is mainlydue to four pigs where the residual after linear regression is outsidethe 95% confidence interval (i.e., pigs 11, 14, 19 and 20, hereaftercalled ‘outliers’) – without those four pigs the R2 increases to0.88. In both cases, overall fit was significant (P < 0.001).To illustrate the correspondence between the video basedobservations and RFID registrations for different pigs and to findan explanation for the outlier pigs, the signals registered for a14 min period containing feeding are illustrated in Fig. 4 for pigs11, 13 and 14 (2 outlier pigs and a pig on the regression line). Itcan be clearly seen that the RFID registrations during feeding forpig 11 are systematically lower than the video registrations, whilethose for pig 13 agree and those for pig 14 are higher. This explainswhy in Fig. 3, pig 11 is below the regression line, pig 13 is on theline and pig 14 is above it. The lower RFID registrations for pig11 and 19 (outliers on Fig. 3) can be explained by these pigs’ unusualนิสัยของมักจะนอนลงในขณะที่ให้อาหาร หมูสองเหล่านี้ได้แสดงพฤติกรรมนี้หลายครั้งในช่วงสุกรโฟกัสเท่านั้นสังเกตวิดีโอ ตำแหน่งนี้มีแนวโน้มเพิ่มขึ้นระยะทางระหว่างแท็ก RFID หูของพวกเขาและเสาอากาศ RFID ดังกล่าวข้างต้นหัว ซึ่งดังนั้น อาจมีลดจำนวนการลงทะเบียนทั้งสี่ของ outliers (หมู 11, 14, 19 และ 20) พบว่าแตกต่างกันอาหารพฤติกรรมของสุกรโฟกัสในระหว่างการสังเกตระยะเวลาการ นี้อธิบายค่าของรอบนอกในแผนการกระจาย(3 รูป) โดยตรวจสอบหมายเลขของการลงทะเบียนของ RFID ต่อวันในระหว่างระยะ fattening สำหรับสุกร 11, 14, 19 และ 20 เมื่อเปรียบเทียบบรรดาหมู 13 เราสามารถ deduce ว่าอาหารนี้พฤติกรรมที่ไม่สอดคล้องกัน หรือไม่ (ไม่แสดงข้อมูล) จำนวนการลงทะเบียนของหมู 11 และ 19 หมูมักจะต่ำกว่าของหมูที่ 13 ขณะที่หมู 14 และ 20 ได้สูงกว่าโดยทั่วไปบรรดาหมู 13 นี้เป็นเฉพาะกรณีในวัน 30วันสังเกต