Area of interest: The trend of current software inevitably leads to the big data era. There are much of large software developed from hundreds to thousands of modules. In software development projects, finding the defect proneness manually on each module in large software dataset is probably inefficient in resources. In this task, the use of a software defect prediction model becomes a popular solution with much more cost-effective rather than manual reviews. This study presents a specific machine learning algorithm, which is the spectral classifier, to develop a software defect prediction model using unsupervised learning approach. Background and objective: The spectral classifier has been successfully used in software defect prediction because of its reliability to consider the similarities between software entities. However, there are conditional issues when it uses the zero value as partitioning threshold. The classifier will produce the predominantly cluster when the eigenvector values are mostly positives. Besides, it will also generate low clusters compactness when the eigenvector contains outliers. The objective of this study is mainly to propose an alternative partitioning threshold in dealing with the zero threshold issues. Generally, the proposed method is expected to improve the spectral classifier based software defect prediction performances. Methods: This study proposes the median absolute deviation threshold based spectral classifier to carry out the zero value threshold issues. The proposed method considers the eigenvector values dispersion measure as the new partitioning threshold, rather than using a central tendency measure (e.g., zero, mean, median). The baseline method of this study is the zero value threshold based spectral classifier. Both methods are performed on the signed Laplacian matrix to meet the non-negative Laplacian graph assumption. For classification, the heuristic row sum method is used to assign the entity class as the prediction label. Results and conclusion: In terms of clustering, the proposed method can produce better cluster memberships that affect the cluster compactness and the classifier performances improvement. The cluster compactness average of both the proposed and baseline methods are 1.4 DBI and 1.8 DBI, respectively. In classification performance, the proposed method performs better accuracy with lower error rates than the baseline method. The proposed method also has high precision but low in the recall, which means that the proposed method can detect the software defect more precisely, although in the small number in detection. The proposed method has the accuracy, precision, recall, and error rates with average values of 0.79, 0.84, 0.72, and 0.21, respectively. While the baseline method has the accuracy, precision, recall, and error rates with average values of 0.74, 0.74, 0.89, and 0.26, respectively. Based on those results, the proposed method able to provide a viable solution to address the zero threshold issues in the spectral classifier. Hence, this study concludes that the use of the median absolute deviation threshold can improve the spectral based unsupervised software defect prediction method.

Imperceptibility and robustness are a part of trade-off components in image watermarking schemes and become one of the most studied in this field. Increasing imperceptibility usually will reduce the robustness, conversely, increasing the robustness may reduce the imperceptibility. This study is intended to overcome this challenge by proposing an image watermarking scheme that increases both imperceptibility and robustness using the Mojette transform. This transform is derived from the Radon transform which has some advantages in geometrical representations that have benefits for imperceptibility and robustness analyses. For evaluation purposes, the performance of the Mojette transform is evaluated on an image watermarking scheme in the singular value decomposition as this domain is stable in response to various image processing operations and also has a simple computation. The use of the Mojette transform in the proposed method involves both main image watermarking processes, which are watermark embedding and extraction. Toward achieving the trade-off of the watermarking requirement, the embedding watermark is designed at the maximum capacity. This means that the watermark is designed at the same size as the host image although it is embedded in various levels of scale factor. The watermarked image is then evaluated in the imperceptibility aspect, and the extracted watermark is evaluated in the robustness aspect after some attacks are applied to thewatermarked images. Experimental results on some test images showthat embedding thewatermark at the maximum size using the proposed method kept producing a good perceptible watermarked image without reducing imperceptibility and robustness. For example, the imperceptibility of the watermarked image at 1% of the embedded scale factor could produce the visual quality at 45.2373 dB and the watermark could extracted well although some attacks have occurred on the watermarked image. The extracted watermark could recovered with almost all of the normalized correlation coefficients above 95% after being distorted by nosing, filtering, compression, rotation, cropping, and flipping attacks. Although the payload capacity of the watermark still needs further analysis, however, this study shows that the Mojette transform can contribute toward achieving the trade-off in the image watermarking requirements.

Stock price fluctuations in the stock market are widely influenced by financial environment changes in both micro and macro that are usually unpredictable and can not be controlled by stock players. On the other side, stock price information is very essential and much needed for both buyers and traders. Stock price forecasting is one of the analytical techniques to obtain stock price prediction based on the previous historical stock prices. The open and close prices are important information in stock trading. The opening price can influence the movement towards the closing price, and the closing price becomes important for the next day's opening price. In technical analysis, the relationship between the two stock prices can be parametric or non-parametric. This study discusses the stock price prediction or forecasting through the non-parametric approach using a multivariate singular spectrum analysis method with the consideration that open and close prices are simultaneously working in the same system and time. Performance evaluation using Mean Absolute Percentage Error shows that the multivariate singular spectrum analysis method can produce predicted open and close prices with an error rate of 3.18% and 3.21%, respectively. Hence, this method can be used as an alternative for stock price forecasting simultaneously.

A novel approach was proposed to develop a composition generation system for creating note sequences representing question-and-answer phrases (QAPs) in melodic phrases. Using a small dataset containing fewer than 10 compositions, the system is expected to generate new melodies that extend beyond the limitations of the original dataset. The modeling of note sequence data and selecting appropriate notes in specific sequences within the melody were chosen for gamelan skeletal melody generation using the Genetic Algorithm (GA) method. Gamelan, a traditional musical ensemble from Java, was selected as the focus of this research, with experiments centered on the gamelan skeletal melody, which plays a role similar to chord progressions in Western music. The evaluation was conducted by statistically analyzing the sequence of notations generated from the dataset and through expert testing. The experimental results demonstrate that the proposed method can generate compositions with well-formed melodic phrases through a small dataset containing varying lengths, data mapping, feature selection based on QAP patterns. GA can enhance creativity in creating note sequences and new QAP patterns, including novel patterns.

Fetal health is very important, it is important for prospective mothers to know from an early age, until now a mother’s sense of ignorance abpout fetal health is very lacking and can result in fetal death. This is due to the lack of curiosity possessed by prospective moyhers and the lack of socialization and infrastructure from related parties about fetal health. Basically, the growth and development of a prospective fetus is very important so that it can be born healthy and without any obstacles at all.The pupose of this study was to detect fetal health using a decision tree classification algorithm with forward feature selection. This experiment was conducted using a public dataset of 2.126 patient data. The results showes that classification using a decision tree algorithm using a decision tree algorithm without feature selection resulted in an accuracy of 89.84%. While the use of the forward selection feature in this decision tree algorithm produces an accuracy of 91.06%.This shows that the use of the forward selection feature can increase accuracy by 1.22%.

Pilkada Serentak is a very important event for the future viability regions and countries. Through this election people can cast their vote and elect representatives of the people according to their choice. Public respond can be expressed through twitter social media. Using twitter social media sentiment analysis can then be made about the public response to the implementation of the election simultaneously. The classification process can be detected via text tweeted by twitter users. In this study, the classification of responses detected by text because it is easily obtained and applied. This study determined the classification of the response to the Indonesian language text and increase accuracy by using SVM.Tweet classification method used by the categorical approach is divided into two classes tweet basic level: positive and negative. Data collected from Indonesian twitter tweet as much as 3000. The labeling is not done manually but using clustering method that divides the 3000 data into two groups. Cluster 1 as a group of positive tweets and Cluster 2 as a negative group tweet.2700 for training data and 300 for the test data. The stage of pre-processing the data includetokenization, casenormalization, stop word detection, and stemming. The process of classification using Support Vector Machine (SVM). Accuracy of SVM showed the highest yield that is 91% compared to the k-means clustering with the results of 82%.