In vivo prediction of abdominal fat and breast muscle in broiler chicken using live body measurements based on machine learning.

The purpose of this study was to predict the carcass characteristics of broilers using support vector regression (SVR) and artificial neural network (ANN) model methods. Data were obtained from 176 yellow feather broilers aged 100-day-old (90 males and 86 females). The input variables were live body measurements, including external measurements and B-ultrasound measurements. The predictors of the model were the weight of abdominal fat and breast muscle in male and female broilers, respectively. After descriptive statistics and correlation analysis, the datasets were randomly divided into train set and test set according to the ratio of 7:3 to establish the model. The results of this study demonstrated that it is feasible to use machine learning methods to predict carcass characteristics of broilers based on live body measurements. Compared with the ANN method, the SVR method achieved better prediction results, for predicting breast muscle (male: R2 = 0.950; female: R2 = 0.955) and abdominal fat (male: R2 = 0.802; female: R2 = 0.944) in the test set. Consequently, the SVR method can be considered to predict breast muscle and abdominal fat of broiler chickens, except for abdominal fat in male broilers. However, further revaluation of the SVR method is suggested.

[Discussion of the Application and Influence Factors of Low-Background Gamma Spectrometry in Distinguishing Sedimentary Rock Types].

The contents of radionuclides uranium, thorium and potassium in the sedimentary rocks mainly depend on the contents of clay in the rocks. And the content of clay is the main basis for distinguishing types of sedimentary rock. Therefore, the value of specific activity or content of uranium, thorium and potassium can be as the quantitative index to distinguish sedimentary rock type. The specific activity or content of radionuclides uranium, thorium and potassium with the method of low-background gamma spectrometry can distinguish the type of rock quickly and accurately. Because of the influence of geometry, mass and moisture content in the sample, the accuracy of distinguishing types of rocks is influenced. This paper makes a theoretical discussion and experimental verification on the influence of mass and moisture content on the results of low-background gamma spectrometry. Results show that there is a linear relationship between (cps) of characteristic peak of all radionuclides and the mass of sample while different energy ranges and lithologies have different linear coefficient and trend fitting degree; The moisture content which is no more than 10%(while collecting samples, the moisture content is no more than 10%) has a little influence on the measurement results( the change values are within the twice standard deviation), so the moisture content which has no significant influence on the accuracy of distinguishing types of sedimentary rock using the method of low-background gamma spectrometry could not be considered. The distinguishing experiment of drilling cuttings samples collected from one oil and gas exploration area in Shanxi Dingbian is done. By the mass correction of the measured data, normalized (cps) ((cps) of per unit mass) of uranium, thorium and potassium channel can only roughly divide the types of sedimentary rocks. Therefore, synthetic distinguishing mode is established with (cps) of combination peak of characteristic peak of uranium, thorium and potassium. The type of rocks is further subdivided, and the distinguishing accuracy is more than 75%.

[Visible/NIR analysis of fat, protein and water in chilled pork].

Fat, protein and water were determined by visible and NIR transmittance spectroscopy in chilled pork. After preprocessed by multiplicative scatter correction (MSC), the quantitative analysis models were developed based on the original, first derivative and second derivative spectra by using partial least squares (PLS) at the temperatures of 0-4 degrees C and 20 degrees C, respectively. By comparing the correlation coefficient (r), RMSEC, and SEP, we found that the first derivative model was the best, and the performance for 0-4 degrees C was better than that for 20 degrees C. At 0-4 degrees C and 20 degrees C, the correlation coefficients were 0.950 and 0.924 for fat, 0.713 and 0.455 for protein and 0.944 and 0.914 for water respectively, SEP values were 2.41 and 2.95 for fat, 5.44 and 4.25 for protein, and 2.37 and 2.38 for water respectively. The results showed that the visible and NIR analysis could measure the fat and water contents in chilled pork well, but was bad for protein, and this was caused by processing line of chilled pork. What's more, the spectrum offset was found in the original spectra at about 770 nm to be about 10 nm.