KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/kmitl/1

Browse

Settings

Subject: search.filters.subject.Extreme gradient boosting

Search Results

Now showing 1 - 2 of 2
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Predictive modeling of bruising in broiler chickens using machine learning algorithms
    (2025-11-01)
    Pirompud, Pranee
    ;
    Sivapirunthep, Panneepa
    ;
    Punyapornwithaya, Veerasak
    ;
    Chaosap, Chanporn
    Bruising chicken broiler is caused by physical stress and injury to the skin and underlying tissues is a major problem in poultry production, affecting both animal welfare and economic outcomes. The aim of this study was to classify the bruising class (low or high percentage of carcass showing bruise at slaughterhouse) per truckload comparing the predictive performance of six machine learning (ML) models– Least Absolute Shrinkage and Selection Operator (LASSO), Classification Tree (CT), Random Forest (RF), Naïve Bayes (NB), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGB)– and using a data set including information about season, time of transport, sex of the flock, flock size, chicken age, chicken mean body weight, housing stocking density, on farm mortality and culling rate, and feed withdrawal time. The general objective was to offer tools for the early detection of flocks with a higher likelihood of bruising and to highlight how ML can support decision-making, strengthen welfare monitoring programs, and reduce economic losses in commercial broiler production. The dataset comprised 26,031 truckloads, each averaging 3,026 ± 245 broilers, and included 14 predictors spanning the rearing, catching, transportation, and slaughtering phases. Data were collected from a commercial broiler producer in Thailand between 2021 and 2022. Bruising percentages ranged from 0 to 4%, with values above 0.8% classified as high, accounting for 41.4% of all truckloads. Due to this distribution, model training was conducted without resampling. Among the evaluated models, the XGB algorithm outperformed the other five machine learning models in predicting bruising in broilers, although the overall predictive accuracy was moderate, likely due to the absence of certain key variables. The most important variables included mean body weight, transport duration, in housing and transport crate stocking density, transport distance, mortality and culling rate throughout the birds' entire life at the farm, feed withdrawal time, and duration in lairage. These findings demonstrate the utility of ML, especially XGB, for supporting early identification of broiler flocks prone to bruising. By applying targeted management interventions—such as maintaining bird health, in housing and transport crate stocking density, feed withdrawal, and lairage duration—producers can mitigate bruising, thereby enhancing both animal welfare and operational efficiency.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    ConvXGB: A new deep learning model for classification problems based on CNN and XGBoost
    (2021-02-01)
    Thongsuwan, Setthanun
    ;
    Jaiyen, Saichon
    ;
    Padcharoen, Anantachai
    ;
    Agarwal, Praveen
    We describe a new deep learning model - Convolutional eXtreme Gradient Boosting (ConvXGB) for classification problems based on convolutional neural nets and Chen et al.’s XGBoost. As well as image data, ConvXGB also supports the general classification problems, with a data preprocessing module. ConvXGB consists of several stacked convolutional layers to learn the features of the input and is able to learn features automatically, followed by XGBoost in the last layer for predicting the class labels. The ConvXGB model is simplified by reducing the number of parameters under appropriate conditions, since it is not necessary re-adjust the weight values in a back propagation cycle. Experiments on several data sets from UCL Repository, including images and general data sets, showed that our model handled the classification problems, for all the tested data sets, slightly better than CNN and XGBoost alone and was sometimes significantly better.