Predict financial distress is an important thing for investor's need to decide whether should put the capital or not. The earning's performance is one of the sign of the distress. Case based reasoning is implementing the highest level of data similarity in its concept. The algorithm should be supported by huge data. This is in line with the Indonesia Stock Exchange's data that has not been maximally utilized. This CBR algorithm can be used to assess the financial company performance in order to make the right investment decision. The ability of this algorithm to do the prediction is showed by the accuracy result of 82%. This research also confirmed that the complete CBR system improves all of the evaluation scoring.
This work evaluates usage feature selection methods to reduce the number of features required to predict docking results between Indonesian medicinal plant compounds and HIV protease. Two feature selection methods, Recursive Feature Elimination (RFE) and Wrapper Method (WM), are trained with a dataset of 7,330 samples and 667 features from PubChem Bioassay and DUD-E decoys. To evaluate the selected features, a dataset of 368 Indonesian herbal chemical compounds labeled by manually docking to PDB HIV-1 protease is used to benchmark the performance of linear SVM classifier using different sets of features. Our experiments show that a set of 471 features selected by RFE and 249 by WM achieve a reduction of classification time by 4.0 and 8.2 seconds respectively. Although the accuracy and sensitivity are also increased by 8% and 16%, no meaningful improvement observed for precision and specificity.
7 April 2023 in Transportation Research Part C: Emerging Technologies
Breast cancer is a serious disease that requires data analysis for diagnosis and treatment. Clustering is a data mining technique that is often used in breast cancer research to assess the level of malignancy at an early stage. However, clustering categorical data can be challenging because different levels in categorical variables can impact the clustering process. This research proposes a modified entropy measure (MEM) to enhance clustering performance. MEM aims to address the issue of distance-based measures in clustering categorical data. It is also a useful tool for assessing data loss in categorical clustering, which helps to understand the patterns and relationships by quantifying the information lost during clustering. An evaluation compares k-modes+MEM, k-means+MEM, DBSCAN+MEM, and affinity+MEM with conventional clustering algorithms. The assessment metrics of clustering accuracy, intra-cluster distance and fowlkes-mallow index (FMI) are employed to evaluate the algorithm performance. Experimental results show significant improvements. k-Modes+MEM algorithm achieves a reduction in average intra-cluster distance and outperforms other algorithms in accuracy, intra-cluster distance, and FMI. The proposed algorithm can be extended to heterogeneous datasets in various domains such as healthcare, finance, and marketing.