Page 180 - Kỷ yếu hội thảo khoa học lần thứ 12 - Công nghệ thông tin và Ứng dụng trong các lĩnh vực (CITA 2023)

164

Equation (7) can be used as a scoring function to measure the quality of a tree
structure q. This score is the same as the classification score for evaluating decision
trees, except that it is computed for a wider range of objective functions.
In fact, the proven XGBoost algorithm optimizes speed and performance for
building predictive models. At the same time, the XGBoost algorithm uses a variety
of data formats, including tabular data of different sizes and layered data types. The
algorithm of XGBoost is now widely by data scientists, is a scalable machine learning
system for tree boosting which can help to avoid over-fitting. It performs well on its
own and has been shown to be successful in many machine learning problems.
In our model, the XGBoost algorithm was implemented as the tree model is usually
used as a primary classifier in XGBoost System. The features extracted from CNN
were fed to train and test the XGBoost classifier in this study.

4 Results

4.1 Dataset

This study uses the dataset name as WEAPD [10] that contained 6,862 images. The
authors have collected weather images from various sources. Figure 3 below depicts a
few examples of images from this dataset.

Fig 3. Example of weather images dataset

CITA 2023 ISBN: 978-604-80-8083-9