KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Search Results

Now showing 1 - 10 of 25

Some of the metrics are blocked by your
consent settings
Unleashing Hidden Business Insights: Harnessing Unstructured Big Data through Text Analysis, NLP, and Visualizations for Budgetary Decisions in Governmental Organizations
(2024-01-01)
Kongthong, Chanwit
;
Jitkajornwanich, Kulsawasd
;
Intakosum, Sarun
Processing Thai language texts can be a challenge due to the complexities of the language, particularly texts from social media and online platforms. This paper introduces an analysis and visualization framework specifically designed to tackle the intricacies associated with processing the Thai language data within the context of online textual content, by utilizing natural language processing (NLP) and visualization techniques. The objectives of this study were to develop an effective Thai text data analysis and visualization framework that allows us to effectively and automatically get a better understanding of the content embedded in Thai textual data. The methodology initiated with a review of existing analysis frameworks and visualization techniques with a specific focus on Thai. The data collection phase encompassed a diverse corpus of Thai text data gathered from online sources. The selected data underwent preprocessing to address language-specific challenges. The proposed Thai analysis and visualization framework consists of multiple stages. Each stage is tailored to accommodate the intricacies of the Thai language, facilitating improved information extraction and text comprehension. The proposed visualization techniques utilize interactive graphs, such as bar charts, line charts, pie charts and donut charts, to offer intuitive and insightful representations of the processed data. Results from our case study show the effectiveness of our Thai analysis framework and visualization techniques in capturing crucial information from online contents written in Thai from governmental organizations.
Some of the metrics are blocked by your
consent settings
Improving OpenAI's Whisper Model for Transcribing Homophones in Legal News
(2024-01-01)
Siriket, Lattapon
;
Jitkajornwanich, Kulsawasd
;
Jaiyen, Saichon
;
Intakosum, Sarun
The 'Whisper' model provides a tool for those who require transcription of human voice. It equips with opensource features and diverse functionalities. The model is capable of effectively deciphering messages in multiple languages, including support for the Thai language. This paper focuses on improving the transcription process of Thai homophones using the Whisper model in reducing the word error rate (WER). We focus on words in the legal news category and identify factors that lead to Whisper's incorrect sound predictions. We examined homophones using snippets of legal news video clips and compiled them into a homophone dictionary. We compare words extracted from the Whisper model by determining the word error rate and spelling of words. Based on the initial results obtained from the original Whisper model and the created homophone dictionary, 48 % of the words were incorrectly transcribed out of a total of 94 words. Then, we propose a methodology by which the performance of the Whisper is improved. That way, the automatic speech recognition of Thai language using the Whisper model can fully be utilized and used in other applications.
Some of the metrics are blocked by your
consent settings
Plant Species Classification Using Leaf Edge Feature Combination with Morphological Transformations and SIFT Key Point
(2023-03-01)
Thomkaew, Jiraporn
;
Intakosum, Sarun
This paper presents a new approach to plant classification by using leaf edge feature combination with Morphological Transformations and defining key points on leaf edge with SIFT. There are three steps in the process. Image preprocessing, feature extraction, and image classification. In the image preprocessing step, image noise is removed with Morphological Transformations and leaf edge detect with Canny Edge Detection. The leaf edge is identified with SIFT, and the plant leaf feature was extracted by CNN according to the proposed method. The plant leaves are then classified by random forest. Experiments were performed on the PlantVillage dataset of 10 classes, 5 classes of healthy leaves, and 5 classes of diseased leaves. The results showed that the proposed method was able to classify plant species more accurately than using features based on leaf shape and texture. The proposed method has an accuracy of 95.62%.
Some of the metrics are blocked by your
consent settings
Automated Data Digitization System for Vehicle Registration Certificates Using Google Cloud Vision API
(2022-07-01)
Thammarak, Karanrat
;
Sirisathitkul, Yaowarat
;
Kongkla, Prateep
;
Intakosum, Sarun
This study aims to develop an automated data digitization system for the Thai vehicle registration certificate. It is the first system developed as a web service Application Programming Interface (API), which is essential for any enterprise to increase its business value. Currently, this system is available on “www.carjaidee.com”. The system involves four steps: 1) an embedded frame aligns a document to be correctly recognised in the image acquisition step; 2) sharpening and brightness filtering techniques to enhance image quality are applied in the pre-processing step; 3) the Google Cloud Vision API receives a prompt to proceed in the recognition step; 4) a specific domain dictionary to improve accuracy rate is developed for the post-processing step. This study defines 92 images for the experiment by counting the correct words and terms from the output. The findings suggest that the proposed method, which had an average accuracy of 93.28%, was significantly more accurate than the original method using only the Google Cloud Vision API. However, the system is limited because the dictionaries cannot automatically recognise a new word. In the future, we will explore solutions to this problem using natural language processing techniques.
Some of the metrics are blocked by your
consent settings
Comparative analysis of Tesseract and Google Cloud Vision for Thai vehicle registration certificate
(2022-04-01)
Thammarak, Karanrat
;
Kongkla, Prateep
;
Sirisathitkul, Yaowarat
;
Intakosum, Sarun
Optical character recognition (OCR) is a technology to digitize a paper-based document to digital form. This research studies the extraction of the characters from a Thai vehicle registration certificate via a Google Cloud Vision API and a Tesseract OCR. The recognition performance of both OCR APIs is also examined. The 84 color image files comprised three image sizes/resolutions and five image characteristics. For suitable image type comparison, the greyscale and binary image are converted from color images. Furthermore, the three pre-processing techniques, sharpening, contrast adjustment, and brightness adjustment, are also applied to enhance the quality of image before applying the two OCR APIs. The recognition performance was evaluated in terms of accuracy and readability. The results showed that the Google Cloud Vision API works well for the Thai vehicle registration certificate with an accuracy of 84.43%, whereas the Tesseract OCR showed an accuracy of 47.02%. The highest accuracy came from the color image with 1024×768 px, 300dpi, and using sharpening and brightness adjustment as pre-processing techniques. In terms of readability, the Google Cloud Vision API has more readability than the Tesseract. The proposed conditions facilitate the possibility of the implementation for Thai vehicle registration certificate recognition system.
Some of the metrics are blocked by your
consent settings
Improvement Classification Approach in Tomato Leaf Disease using Modified Visual Geometry Group (VGG)-InceptionV3
(2022-01-01)
Thomkaew, Jiraporn
;
Intakosum, Sarun
This paper presents a new method for optimizing tomato leaf disease classification using Modified Visual Geometry Group (VGG)-InceptionV3. Improved performance of VGG-16 model as a base model with InceptionV3 block reduced the number of convolution layers of VGG-16 from 16 to 10 layers, and added an InceptionV3 block that was improved by adding convolution layer from 3 to 4 layers to increase the accuracy of tomato leaf disease classification and reduce the number of parameters and computation time of the model. The experiments were performed on tomato leaves from the PlantVillage dataset of 10 classes, consisting of nine classes of diseased leaves and one class of healthy leaves. The results showed that the proposed method was able to reduce the number of parameters and computation time with and accuracy of tomato leaf disease classification was 99.27%. Additionally, the proposed approach was compared with state-of-the-art Convolutional Neural Network (CNN) models such as VGG16, InceptionV3, DenseNet121, MobileNetV2, and RestNet50. Comparative results showed that the proposed method had the highest accuracy in the tomato leaf disease classification and required a smaller number of parameters and computational time
Some of the metrics are blocked by your
consent settings
Discovery of association rules between factors affecting user satisfaction in software project by FP-GROWTH Algorithm
(2021-05-19)
Kaewbanjong, Katawut
;
Intakosum, Sarun
The main objective of this work was to find the association rules between factors affecting user satisfaction in software project by using an association rule discovery technique. Data from 191 software projects were collected and association rules between 15 Selected factors were discovered by FP-growth algorithm. Primarily, 281 association rules were discovered. Rules that were not directly related to user satisfaction were filtered out by simple criteria. In the end, 11 final association rules that passed those criteria were obtained with an average confidence value of 65.30%. These rules can incorporated in the planning and administration of a software project to gain utmost user satisfaction
Some of the metrics are blocked by your
consent settings
Statistical Analysis with Prediction Models of User Satisfaction in Software Project Factors
(2020-06-01)
Kaewbanjong, Katawut
;
Intakosum, Sarun
We analyzed a volume of software project data and found significant user satisfaction in several software project factors. statistical significance A analysis (logistic regression) a collinearity analysis and determined the significance factors from a group of 71 pre-defined factors from 191 software projects in ISBSG Release 12. Eight prediction models were used to test the prediction potential of these factors: Neural network, k-NN, Naïve Bayes, Random forest, Decision tree, Gradient boosted tree, linear regression and logistic regression prediction model. Fifteen pre-defined factors were significant in predicting user satisfaction: client-server, personnel changes, total defects delivered, project inactive time, industry sector, application type, development type, how methodology was acquired, development techniques, decision making process, intended market, size estimate approach, size estimate method, cost recording method, and effort estimate method. They provided 82.71% prediction accuracy when used with a neural network prediction model. These findings may directly benefit software development managers.
Some of the metrics are blocked by your
consent settings
High candidates generation: A new efficient method for mining share-frequent patterns
(2017-11-01)
Nawapornanan, Chayanan
;
Intakosum, Sarun
;
Boonjing, Veera
The share frequent patterns mining is more practical than the traditional frequent patternset mining because it can reflect useful knowledge such as total costs and profits of patterns. Mining share-frequent patterns becomes one of the most important research issue in the data mining. However, previous algorithms extract a large number of candidate and spend a lot of time to generate and test a large number of useless candidate in the mining process. This paper proposes a new efficient method for discovering share-frequent patterns. The new method reduces a number of candidates by generating candidates from only high transaction-measure-value patterns. The downward closure property of transaction-measure-value patterns assures correctness of the proposed method. Experimental results on dense and sparse datasets show that the proposed method is very efficient in terms of execution time. Also, it decreases the number of generated useless candidates in the mining process by at least 70%.
Some of the metrics are blocked by your
consent settings
Borderline over-sampling in feature space for learning algorithms in imbalanced data environments
(2016-01-01)
Savetratanakaree, Kittipat
;
Sookhanaphibarn, Kingkarn
;
Intakosum, Sarun
;
Thawonmas, Ruck
In this paper, we propose a new approach to over-sample new minority-class instances along the borderline using the Euclidean distance in the feature space to improve support vector machine (SVM) performance in imbalanced data environments. SVM has been an outstandingly successful classifier in a wide variety of applications where balanced class data distribution is assumed. However, SVM is ineffective when coping with imbalanced datasets whereby the majorityclass instances far outnumber the minority-class instances. Our new approach, called Borderline Over-sampling in the Feature Space, can deal with imbalanced data to effectively recognize new minority-class instances for better classification with SVM. The results of our class prediction experiments using the proposed approach demonstrate better performance than the existing SMOTE, Borderline-SMOTE and borderline over-sampling methods in terms of the g-mean and F-measure.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results