Thusaranon, Panita

Loading...
Thumbnail Image
Preferred name
Thusaranon, Panita
Email
panita.th@kmitl.ac.th

Settings

Now showing 1 - 2 of 2
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Frequent itemsets mining using random walks for record insertion and deletion
    (2017-02-23)
    Thusaranon, Panita
    ;
    Kreesuradej, Worapoj
    In Association rules mining, the task of finding frequent itemsets in dynamic database is very important because the updates may not only invalidate some existing rules but also make other rules relevant. In this paper, we propose a new algorithm to maintain frequent itemsets of a dynamic database in the case of record insertion as well as deletion simultaneously. Basically, the proposed algorithm maintains not only the support counts of frequent itemsets but also the support counts of prospective frequent itemsets, i.e., infrequent itemsets that promise to be frequent in the future, in an original database. Prospective frequent itemsets, which are obtained by using the principle of Random Walks, can help to reduce a number of times to rescan the original database.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    SmartAir: Enhancing Air Quality Classification with Deep Learning and Two-State Q-Learning
    (2025-01-01)
    Peung-Uaypon, Supitchaya
    ;
    Jitkongchuen, Duangjai
    ;
    Thusaranon, Panita
    Air pollution, especially PM 2.5, poses a growing global health threat, mainly from industrial activity, traffic, and wildfires. Traditional air quality monitoring systems are costly and lack broad coverage. This research proposes a low-cost alternative using a model that classifies air quality from outdoor images by combining Deep Learning with Reinforcement Learning. The model uses VGG19 (pre-trained on ImageNet) along with Local Binary Pattern (LBP) and RGB average values for feature extraction. A non-linear SVM classifier is enhanced with Q-Learning, which improves classification of difficult images through randomized actions: rotating images ± 15 degrees or shifting them diagonally. Experimental results show that incorporating Q-Learning increased model accuracy from 96.11% to 97.29%. This indicates that reinforcement learning helps the system adaptively correct misclassifications. The proposed model offers an efficient, accessible, and affordable tool for air quality monitoring, especially in areas lacking conventional AQI sensors.