Using Rough Set Theory to Find Minimal Log with Rule Generation

Data pre-processing is a major difficulty in the knowledge discovery process, especially feature selection on a large amount of data. In literature, various approaches have been suggested to overcome this difficulty. Unlike most approaches, Rough Set Theory (RST) can discover data de-pendency and reduce the attributes without the need for further information. In RST, the discernibility matrix is the mathematical foundation for computing such reducts. Although it proved its efficiency in feature selection, unfortunately it is computationally expensive on high dimensional data. Algorithm complexity is related to the search of the minimal subset of attributes, which requires computing an exponential number of possible subsets. To overcome this limitation, many RST enhancements have been proposed. Contrary to recent methods, this paper implements RST concepts in an iterated manner using R language. First, the dataset was partitioned into a smaller number of subsets and each subset processed independently to generate its own minimal attribute set. Within the iterations, only minimal elements in the discernibility matrix were considered. Finally, the iterated outputs were compared, and those common among all reducts formed the minimal one (Core attributes). A comparison with another novel proposed algorithm using three benchmark datasets was performed. The proposed approach showed its efficiency in calculating the same minimal attribute sets with less execution time.

Attribute selection approaches for incomplete interval-value data

Attribute selection in an information system (IS) is an important issue when dealing with a large amount of data. An IS with incomplete interval-value data is called an incomplete interval-valued information system (IIVIS). This paper proposes attribute selection approaches for an IIVIS. Firstly, the similarity degree between two information values of a given attribute in an IIVIS is proposed. Then, the tolerance relation on the object set with respect to a given attribute subset is obtained. Next, θ-reduction in an IIVIS is studied. What is more, connections between the proposed reduction and information entropy are revealed. Lastly, three reduction algorithms base on θ-discernibility matrix, θ-information entropy and θ-significance in an IIVIS are given.

Attribute Reduction Method Based on Sample Extraction and Priority

Attribute reduction is a key issue in the research of rough sets. Aiming at the shortcoming of attribute reduction algorithm based on discernibility matrix, an attribute reduction method based on sample extraction and priority is presented. Firstly, equivalence classes are divided using quick sort for computing compressed decision table. Secondly, important samples are extracted from compressed decision table using iterative self-organizing data analysis technique algorithm(ISODATA). Finally, attribute reduction of sample decision table is conducted based on the concept of priority. Experimental results show that the attribute reduction method based on sample extraction and priority can significantly reduce the overall execution time and improve the reduction efficiency.

Fuzzy-Discernibility Matrix-based an Efficient Feature Selection Technique for Improved Motor-Imagery EEG Signal Classification

Brain activities, called brain rhythms, are the micro-level electrical signals (that is, Electroencephalogram or EEG) generated in our brain while we are performing a task. Even when we imagine a limb movement, it generates the same EEG signals called motor-imagery. Motor-imagery-based Brain-computer Interface (BCI) provides a non-muscular means to connect the human brain with limbs through computer-based interpretations. The main aim of this paper is to find a suitable feature-set and a classifier to efficiently classify EEG signals into distinct motor-imagery brain-states. We propose to use sliding temporal window-based approaches for feature extraction from EEG and a mix-bagging classifier which is essentially a bagging-based ensemble of multiple types of learners for motor imagery EEG classification. We observe that mix-bagging with overlapping sliding window-based feature extraction achieves an accuracy of 91.43% on the BCI Competition II Dataset III. To reduce the feature size further, we use a fuzzy discernibility matrix that selects the most discriminative features instead of all the features. This additional feature selection strategy improves the classification accuracy to 92.14% and sets a new state-of-the-art result on this dataset.

Attribute reduction in fuzzy multi-covering decision systems via observational- consistency and fuzzy discernibility

Fuzzy covering rough set model is a popular and important theoretical tool for computation of uncertainty, and provides an effective approach for attribute reduction. However, attribute reductions derived directly from fuzzy lower or upper approximations actually still occupy large of redundant information, which leads to a lower ratio of attribute-reduced. This paper introduces a kind of parametric observation sets on the approximations, and further proposes so called parametric observational-consistency, which is applied to attribute reduction in fuzzy multi-covering decision systems. Then the related discernibility matrix is developed to provide a way of attribute reduction. In addition, for multiple observational parameters, this article also introduces a recursive method to gradually construct the multiple discernibility matrix by composing the refined discernibility matrix and incremental discernibility matrix based on previous ones. In such case, an attribute reduction algorithm is proposed. Finally, experiments are used to demonstrate the feasibility and effectiveness of our proposed method.

Data Mining dalam Akurasi Tingkat Kelayakan Pakai terhadap Peralatan Perangkat Keras

Hardware is a computer that can be seen and touched in person. Hardware is used to support student work and learning processes. The hardware should always be in good shape. If any damage should be done quickly. The benefits of this study provide a viable level of data against hardware tools. The purpose of this study determines that hardware that is worth using quickly and precisely so easily can be repaired and replaced. Hard-processed action consists of 12 projectors, 2 units of access point, 6 units of monitors, and 20 CPU units. To see the level of appropriateness regarding hard drives requires a rough set algorithm with that stage: information system; Decision system; Equivalency class; Discernibility matrix; Discernibility Matrix module D; Reduction; Generate Rules. The results of the 40 devices of study STMIK Indonesia Padang subtract college have 10 rules of policy on whether the hardware is still viable, repaired or replaced. So using a rough set algorithm is particularly appropriate to apply in a verifiable level of accuracy to fast and precise hardware.

Penentuan Tingkat Kerusakan Peralatan Labor Komputer Menggunakan Data Mining Rough Set

Equitments of computer laboratory have a function as an important tools in supporting pratical lecturing. These facilities should always be on a condition like ready are proper to use both computers and others. To avoid equipment detriment, it is necessary to do early identification in which prevent the worse condition of equitments. The method use in this study is rough set method wich consists several stages such as Decision System, Equivalence Class, Discernibility Matrix, Discernibility Matrix Modulo D, Reduction, and Generate Rules. From this study, it was found that 14 rules in making decisions for equipments treatment of computer laboratory such as use, repair and replace. Thus, this mrthod is very capable in determining the detriment level of laboratory equipment.