This study presents a novel data preprocessing framework to enhance pig behavior analysis using machine learning techniques. We address the critical issue of data leakage in time series data, which can lead to overfitting and poor generalization in real-world applications. Our approach introduces two key innovations: a non-class-based windowing method and a chronological time sampling technique for train/test splitting. To evaluate these methods, we collected a comprehensive dataset spanning 100 hours of pig behavior over 24 days, using ear-tag sensors and video recordings to capture 12 distinct activities. We evaluate the effectiveness of our preprocessing methods using various machine learning and deep learning models on both time-domain and feature-domain datasets derived from this unique collection. Results demonstrate significant improvements in classification accuracy across all tested models, with increases of up to 15% compared to commonly used data preparation methods. The 2D Residual CNN achieved the highest accuracy of 95.6% in the time domain, while Random Forest performed best in the feature domain with 94.1% accuracy.

Accurate weight forecasting is essential for optimizing swine farming operations and enhancing animal welfare. This paper introduces a novel approach for pig weight forecasting, employing multi-input deep learning models that harness both depth images and statistical descriptors. The study conducts a comprehensive comparison of traditional machine learning (ML) models, deep learning (DL) models, hybrid ML and DL models, and multi-input models integrating both time-series data and image features. A meticulously curated dataset comprising time-series weight measurements and corresponding depth images of pigs forms the foundation of the study. Image descriptors such as length, width, depth, and volume were extracted from the depth images. The proposed multi-input models, employing architectures based on ResNet, XCeption, LSTM, and GRU layers, are meticulously trained and evaluated using this dataset. The performance evaluation is conducted using mean absolute error (MAE) and mean absolute percentage error (MAPE) metrics. The results underscore the superiority of the multi-input models over traditional ML, DL, and hybrid models. Notably, the best-performing model achieves a test MAE of 1.81 kg and a test MAPE of 5.56 %. This exceptional performance highlights the importance of leveraging both time-series data and image features for precise weight forecasting in pigs. These findings can hold significant implications for improving the efficiency and sustainability of swine farming practices, offering a pathway towards improved decision-making and animal management protocols.

Edge intelligence can reduce power dissipation to enable power-hungry long-range wireless applications. This work applies edge intelligence to quantify the reduction in power dissipation. We designed a wireless sensor node with a LoRa radio and implemented a decision tree classifier, in situ, to classify behaviors of cattle. We estimate that employing edge intelligence on our wireless sensor node reduces its average power dissipation by up to a factor of 50, from 20.10 mW to 0.41 mW. We also observe that edge intelligence increases the link budget without significantly affecting average power dissipation.

This paper investigates direct and indirect learning methods to develop deep learning digital predistortion (DL-DPD) models and apply the models to improve the linearity of a power amplifier (PA). The two methods are applied to class-AB and class-$\mathrm{F}^{-1}$ PAs designed with gallium nitride (GaN) on silicon carbide (SiC) high electron mobility transistors (HEMTs). The simulation results show that both direct and indirect DL-DPD methods improve the linearity of the class-AB PA by about 12dB and the class-$\mathrm{F}^{-1}$ PA by 11 dB, while the indirect method offers marginally better performance. The paper shows the direct learning method leads to significant improvement of the DL-DPD method based over the memory polynomial. It also presents the advantages of a BiLSTM based on the neural network architecture to design direct/indirect DPDs. Finally, it demonstrates that the DL-DPD can improve the linearity of class-AB and class-$F^{-1}$ PAs without architectural changes.

UNLABELLED: Phenotypic Up-regulated Gene Support Vector Machine (PUGSVM) is a cancer Biomedical Informatics Grid (caBIG™) analytical tool for multiclass gene selection and classification. PUGSVM addresses the problem of imbalanced class separability, small sample size and high gene space dimensionality, where multiclass gene markers are defined by the union of one-versus-everyone phenotypic upregulated genes, and used by a well-matched one-versus-rest support vector machine. PUGSVM provides a simple yet more accurate strategy to identify statistically reproducible mechanistic marker genes for characterization of heterogeneous diseases. AVAILABILITY: http://www.cbil.ece.vt.edu/caBIG-PUGSVM.htm.

To construct biologically interpretable features and facilitate Muscular Dystrophy (MD) sub-types classification, we propose a novel integrative scheme utilizing PPI network, functional gene sets information, and mRNA profiling. The workflow of the proposed scheme includes three major steps: First, by combining protein-protein interaction network structure and gene co-expression relationship into new distance metric, we apply affinity propagation clustering to build gene sub-networks. Secondly, we further incorporate functional gene sets knowledge to complement the physical interaction information. Finally, based on constructed sub-network and gene set features, we apply multi-class support vector machine (MSVM) for MD sub-type classification, and highlight the biomarkers contributing to the sub-type prediction. The experimental results show that our scheme could construct sub-networks that are more relevant to MD than those constructed by conventional approach. Furthermore, our integrative strategy substantially improved the prediction accuracy, especially for those hard-to-classify sub-types.

In this paper, we propose to apply non-uniform weighs to the genes in the pathway based analysis and present two weighting schemes for the genes. Specifically, we incorporate our weighting schemes into the global test pathway based analysis approach and investigate the effects of our weighting schemes. We observe that when non-uniform weights are applied, some originally lower ranked pathways are elevated to the high ranks, prediction performances of the selected genes are improved, and some genes associated with the related phenotype are identified, which are missed by the uniform weighting approach.