Effect of De-noising by Wavelet Filtering and Data Augmentation by Borderline SMOTE on the Classification of Imbalanced Datasets of Pig Behavior

Min Jin,Chunguang Wang,Dan Børge Jensen

doi:10.3389/fanim.2021.666855

Min Jin, Chunguang Wang + Show 1 more

Open Access

https://doi.org/10.3389/fanim.2021.666855

Copy DOI

Abstract

Classification of imbalanced datasets of animal behavior has been one of the top challenges in the field of animal science. An imbalanced dataset will lead many classification algorithms to being less effective and result in a higher misclassification rate for the minority classes. The aim of this study was to assess a method for addressing the problem of imbalanced datasets of pigs' behavior by using an over-sampling method, namely Borderline-SMOTE. The pigs' activity was measured using a triaxial accelerometer, which was mounted on the back of the pigs. Wavelet filtering and Borderline-SMOTE were both applied as methods to pre-process the dataset. A multilayer feed-forward neural network was trained and validated with 21 input features to classify four pig activities: lying, standing, walking, and exploring. The results showed that wavelet filtering and Borderline-SMOTE both lead to improved performance. Furthermore, Borderline-SMOTE yielded greater improvements in classification performance than an alternative method for balancing the training data, namely random under-sampling, which is commonly used in animal science research. However, the overall performance was not adequate to satisfy the research needs in this field and to address the common but urgent problem of imbalanced behavior dataset.

Highlights

Behavior is one of the most used and sensitive indicators which can reflect livestock’s physical, physiological, and health status, as well as their reactions to the environment
For the raw accelerometer data to become useful, it must undergo various steps of data preparations. One such step is often de-noising; the random noise generated from the physiological jitter of the animal, or the device being randomly shaken or impacted, will be added to the systematic signals that correspond to the various behaviors, reducing the signal-to-noise ratio of the raw data, and affecting the accuracy of the behavior classification typically performed in later steps
The reason why under-sampling is often preferred to oversampling is likely related to the risk that the simplest form of oversampling, i.e., over-sampling by simple random resampling, could lead to over-fitting

Summary

Introduction

Behavior is one of the most used and sensitive indicators which can reflect livestock’s physical, physiological, and health status, as well as their reactions to the environment. Tri-axial accelerometers, fixed to the bodies of animals, have previously been applied for the collection of data used for animal behavior classification. For the raw accelerometer data to become useful, it must undergo various steps of data preparations. One such step is often de-noising; the random noise generated from the physiological jitter of the animal, or the device being randomly shaken or impacted, will be added to the systematic signals that correspond to the various behaviors, reducing the signal-to-noise ratio of the raw data, and affecting the accuracy of the behavior classification typically performed in later steps

Objectives

Methods

Results

Conclusion