KMITL

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

Browse

Search Results

Now showing 1 - 9 of 9

Some of the metrics are blocked by your
consent settings
DEEP LEARNING DEVELOPMENTAL MODEL FOR EMBRYO GRADING IN THAI POPULATION
(2024-01-01)
Wiriyasirivaj, Budsaba
;
Limkiatsataporn, Sawit
;
Pukinghin, Apisit
;
Kuekulkomain, Phatrapron
;
Promrungrueng, Pornprom
In light of the growing challenges associated with infertility, an increasing number of researchers are resorting to assisted reproductive technologies such as In Vitro Fertilization (IVF). Embryo grading is a crucial step in the IVF process that requires embryologists’ expertise. However, their limited availability has led to the exploration of technological alternatives. This study aims to use deep learning for human embryo grading, with models specifically designed for the dataset in Thailand at Vajira Hospital. The process of IVF at Vajira Hospital presents its own set of challenges since its embryo classification extends beyond the Istanbul consensus. Furthermore, classes that occur infrequently are removed and classes with similarities are merged. We apply transfer learning to pretrained deep learning models like VGG19, VGG16, ResNet152, Resnet101, Xception, InceptionV3, and EfficientNet, to create a system capable of accurately classifying embryo quality scores. Experimental results from the embryo dataset collected from Vajira Hospital in Thailand demonstrate the proposed classifier’s accuracy, precision, recall, f1-score, and AUC superiority. This research contributes to the field of IVF in Thailand by potentially reducing human errors and addressing the demand for skilled embryologists.
Some of the metrics are blocked by your
consent settings
Improving knn algorithm based on weighted attributes by pearson correlation coefficient and pso fine Tuning
(2020-10-21)
Sinhashthita, Wanarase
;
Jearanaitanakij, Kietikul
Assigning proper weights to attributes in some datasets according to their importances can significantly improve the classification accuracy. Weighted attributes can support the classification methods effectively if their weights truly represent by their importances. In this research, we improve the K-Nearest Neighbors (KNN) algorithm by using Pearson correlation coefficient along with Particle Swarm Optimization (PSO) to find the optimal set of weights for attributes in the dataset. The experimental results show that the proposed method can significantly improve the classification accuracy when compared to the traditional KNN algorithm.
Some of the metrics are blocked by your
consent settings
Improving ID3 Algorithm by Combining Values from Equally Important Attributes
(2018-08-21)
Kraidech, Suratchanan
;
Jearanaitanakij, Kietikul
ID3 is a well-known algorithm which is used in the classification task of the decision tree learning. Although a lot of research provides improvements on the traditional ID3 algorithm with various strategies, no attempt was made on ID3 to address the problem when there are more than one attribute that can be placed at a particular node, i.e. those attributes are equally important. This paper proposes a new variation of ID3 to combine equally important attributes into a single node of the decision tree classification. The Connect-4 dataset from UCI is used in our experiment since the dataset contains many attributes and instances which can easily encounter the equally important attribute problem. The experimental results show that our proposed method significantly reduces the average depth of the decision tree generated by ID3 algorithm while the average accuracy rate is still preserved.
Some of the metrics are blocked by your
consent settings
Improving ID3 Algorithm by Using A∗ Search
(2018-08-21)
Kaewrod, Nicha
;
Jearanaitanakij, Kietikul
ID3 is one of the most widely used algorithms for creating a classification decision tree. However, the traditional ID3 algorithm has a difficulty when there are equally important attributes during the decision node construction. It randomly selects one of the most important attributes to serve as the current decision node. This behavior may lead to the resulting decision tree which contains unnecessary depths and suboptimal accuracy. The purpose of this paper is to find the near optimal depth decision tree of the dataset, which contains an equally important attributes problem, by using the A-Star (A∗) search. The experiment results on four standard datasets from UCI indicate that the proposed algorithm can significantly reduce the decision tree's depth and still maintain the classification accuracy.
Some of the metrics are blocked by your
consent settings
Reducing the depth of ID3 algorithm by combining values from neighboring important attributes
(2018-07-02)
Kraidech, Suratchanan
;
Jearanaitanakij, Kietikul
The ID3 algorithm is one of the most popular decision tree algorithms which is mainly used in the classification task. There are many pieces of research about improving the ID3 algorithm by using various strategies. We improved the ID3 algorithm by addressing the equally important attributes problem in our previous work. In this paper, we extend our previous algorithm by changing the attribute selection to allow the neighboring second-place important attributes to be combined with the most important attributes. The proposed algorithm is tested on four standard benchmarks from the UCI repository. The experimental results indicate the significant reduction in the maximum depth and the classification depth of the decision tree. In addition, the testing time is also reduced while the classification accuracy is satisfying stable.
Some of the metrics are blocked by your
consent settings
Improving ID3 algorithm by ignoring minor instances
(2018-07-02)
Kaewrod, Nicha
;
Jearanaitanakij, Kietikul
Among various classification algorithms, ID3 is one of the most widely used and well-known tools that generates an efficient decision tree. Nevertheless, ID3 is too rigorous in generating the decision rules. As a result, the final decision tree may carry too many decision rules. Some of these decision rules may have very low number of instances which do not make significant change to the classification accuracy. The aim of this paper is to propose an approach to relax the rigorousness of the conventional ID3 algorithm by ignoring minor instances so that the resulting decision tree will have the lower number of depths yet produce promising accuracy. The proposed algorithm is examined on six datasets from UCI repository and Weka. The experimental results indicate that the proposed algorithm not only significantly reduces the maximum number of depths of the decision tree, but also retains the classification accuracy in the satisfying level. Moreover, the training time, the classification time, and the number of decision rules of the proposed algorithm are lower than those of the conventional ID3.
Some of the metrics are blocked by your
consent settings
Spartan simplicity: A pruning algorithm for neural nets
(2008-08-01)
Jearanaitanakij, Kietikul
;
Pinngern, Ouen
Having more hidden units than necessary can produce a neural network that has a poor generalization. This paper proposes a new algorithm for pruning unnecessary hidden units away from the single-hidden layer feedforward neural networks, resulting in a Spartan network. Our approach is simple and easy to implement, yet produces a very good result. The idea is to train the network until it begins to lose its generalization. Then the algorithm measures the sensitivity and automatically prunes away the most irrelevant unit. We define this sensitivity as the absolute difference between the desirable output and the output of the pruned network. Unlike other pruning methods, our algorithm is distinct in calculating the sensitivity from the validation set, instead of the training set, without increasing the asymptotic time complexity of the back-propagation algorithm. In addition, for a classification problem, we raise a point that the sensitivities of some well-known pruning algorithms may still underestimate the irrelevance of hidden unit even though the validation set is used in measuring the sensitivity. We resolve this problem by considering the number of misclassified patterns as the main concern. The Spartan simplicity algorithm is applied to three artificial and seven standard benchmarks. In most problems, the algorithm can produce a compact-sized network with high generalization ability in comparison with other pruning algorithms. © 2008 World Scientific Publishing Company.
Some of the metrics are blocked by your
consent settings
Classifying continuous data set by ID3 algorithm
(2005-12-01)
Jearanaitanakij, Kietikul
This paper presents a modified version of the ID3 algorithm. The goal is to build the decision tree for classifying the continuous data set. An example in the training data set composes of some input features (attributes) and one predicate output. A proper feature ordering produces a shallow decision tree, which spends a logarithm time in classifying a data set. The original ID3 algorithm calculates the information gains of the features and arranges those features by descending order of the information gains. As a result, the decision tree selects a feature which has the biggest information gain at the top level. The algorithm repeats the feature ordering process for the rest of the features until there is not any unclassified example in the training data. However, one problem of the original ID3 algorithm is that it cannot classify the continuous feature in the data set. In order to serve a continuous feature, the ID3 algorithm is modified by quantizing the continuous feature into intervals and performing the classification process within those intervals. The modified algorithm is tested with a standard data set. The experimental results show a relationship between the number of intervals and the error rate on a standard real-world problem. © 2005 IEEE.
Some of the metrics are blocked by your
consent settings
An information gain technique for acceleration of convergence of artificial neural networks
(2005-12-01)
Jearanaitanakij, Kietikul
;
Pinngern, Ouen
This paper presents an application of information gain to accelerate the convergence time of Artificial Neural Networks (ANNs). We improve Hagiwara's convergence acceleration algorithm by applying information gain to it. The first step of our proposed technique is to calculate information gains of all features (or attributes) in training data and pass those gains through all hidden units in the next layer. During the training process, the algorithm monitors sum-squared error at the output layer. When the variation of sum-squared error becomes small, the worst hidden unit is detected. Next, all the weights connected to the worst hidden unit are reset to random values within the appropriate ranges. These ranges are determined by the propagated information gain of the worst hidden unit. Then, the network is retrained. When the number of weight resetting trials reaches a certain number, a new hidden unit is added to the network and the whole training process is repeated. Our experimental results on standard benchmarks show remarkable outputs in terms of convergence time. © 2005 IEEE.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results