Average Accuracy Results Research Articles

Motor Imagery Classification Improvement of Two-Class Data with Covariance Decentering Eigenface Analysis for Brain–Computer Interface Systems

This study is intended to improve the motor imagery classification performance of two-class data points using newly developed covariance decentering eigenface analysis (CDC-EFA). When extracting the classification for the given data points, it is necessary to precisely distinguish the classes because the left and right features are difficult to differentiate. However, when centering is performed, the unique average data of each feature are lost, making them difficult to distinguish. CDC-EFA reverses the centering method to enhance data characteristics, making it possible to assign weights to data with a high correlation with other data. In experiments with the BCI dataset, the proposed CDC-EFA method was used after preprocessing by filtering and selecting the electroencephalogram data. The decentering process was then performed on the covariance matrix calculated when acquiring the unique face. Subsequently, we verified the classification improvement performance via simulations using several BCI competition datasets. Several signal processing methods were applied to compare the accuracy results of the motor imagery classification. The proposed CDC-EFA method yielded an average accuracy result of 98.89%. Thus, it showed improved accuracy compared with the other methods and stable performance with a low standard deviation.

Comparison Of The C.45 And Naive Bayes Algorithms To Predict Diabetes

Diabetes mellitus is an urgent global health problem and has a major impact on people around the world. This disease is characterized by high levels of sugar (glucose) in the blood due to disturbances in the production or use of the hormone insulin by the body. This study aims to carry out accurate early detection of diabetics so that they can be treated as soon as possible to reduce the risk of death and to compare the two algorithms that have the best level of accuracy. The algorithms used in this study are the C4.5 and Naïve Bayes Decision Tree Algorithms. The results of the experiments carried out in this study the Decision Tree Algorithm C4.5 and Naïve Bayes can be used in modeling the early detection of diabetes. The highest average accuracy results were obtained at 90.835% using the Decision Tree C4.5 Algorithm. As for the Naïve Bayes Algorithm, an average accuracy rate of 90.745% is obtained. The pruning process was carried out using the Decision Tree Algorithm C4.5, the accuracy performance increased to 91.30%. There were 18 patterns or rules for the early detection of diabetics from the built model. The determination of attributes, the number of attribute dimensions, and the number of samples greatly affect the performance of the model built.

Detection of multiple anomalous instances in video surveillance systems

Anomalous event recognition has a complicated definition in the complex background due to the sparse occurrence of anomalies. In this paper, we form a framework for classifying multiple anomalies present in video frames that happen in a context such as the sudden moment of people in various directions and anomalous vehicles in the pedestrian park. An attention U-net model on video frames is utilized to create a binary segmented anomalous image that classifies each anomalous object in the video. White pixels indicate the anomaly, and black pixels serve as the background image. For better segmentation, we have assigned a border to every anomalous object in a binary image. Further to distinguish each anomaly a watershed algorithm is utilized that develops multi-level gray image masks for every anomalous class. This forms a multi-class problem, where each anomalous instance is represented by a different gray color level. We use pixel values, Optical Intensity, entropy values, and Gaussian filter with sigma 5, and 7 to form a feature extraction module for training video images along with their multi-instance gray-level masks. Pixel-level localization and identification of unusual items are done using the feature vectors acquired from the feature extraction module and multi-class stack classifier model. The proposed methodology is evaluated on UCSD Ped1, Ped2 and UMN datasets that obtain pixel-level average accuracy results of 81.15%,87.26% and 82.67% respectively.

Jump forecasting in foreign exchange markets: A high‐frequency analysis

AbstractUsing tick data for 14 emerging and developed market currencies covering the period from January 2018 until April 2021, we first detect jumps by Lee and Mykland methodology then apply various machine learning algorithms to forecast out of sample jump occurrences and their direction. Our results show that the arrival and the direction of intraday jumps in the foreign exchange market can be predicted with these algorithms combined with liquidity metrics and technical indicators, even for the Covid pandemic period where volatility in the foreign exchange market is very high. Among all the methods considered, multilayer perceptron has the highest average accuracy for jump prediction overall, followed by support vector machine and random forest methodologies with slightly less average accuracy results. Results are robust to alternative sampling schemes. Accordingly, central bankers can adjust liquidity injection timing with these jump prediction models in the foreign exchange markets where they can try to minimize jump strength if not completely eliminate its occurrence. For investors, having information regarding jump occurrence timings gives an opportunity to hedge against foreign exchange risks more efficiently.

Aspect Based Opinion Summarization Using Rule-Based Method and Support Vector Machine for Indonesian Reviews

E-commerce's reviews feature can help users find information about the desired skin care products to choose the right one. However, the reviews number of a product grows rapidly due to the popularity of e-commerce and the product itself. A user becomes difficult to read all reviews one by one and extract useful information. To deal with this problem, we summarize aspects using the Rule-Based method and Support Vector Machine. We propose a Rule-Based method that is used to break down a review into several segments based on its aspect. Support Vector Machine is used to classify sentence segments according to their polarity. The data used in this study is Indonesian reviews of skin care products obtained from the Female Daily website. The average accuracy results using 10-fold cross-validation of sentiment classification is 74%. We experimented on 462 reviews where the accuracy is 92% in aspect categorization and 71.2% in sentiment classification. Based on humans, the lowest value is the suitability of a sentence with its sentiment/polarity. The highest value is the suitability of sentences with its aspect and usability of summary to helps users to find specific information so they can decide whether to buy that product or not. It can be concluded that the reader can well receive the summary. Future work can consider the negation word to reduce misclassification in the sentiment classification step.

Analisis Sentimen Kata Anjay pada Media Sosial Twitter Dalam Kajian Linguistik Komputasi

Language serves as a means of supporting activities, ideas, and human behavior. Every form of human life can be expressed in language. Language in everyday use depends on the user and the situation of use. Because of these two things, various variations of language are used in communication. In mid-2020, Indonesia was shocked by the ban on the use of the word anjay. The purpose of this study was to determine the classification of text mining methods in filtering the word anjay, to find out how the sentence structure is in each tweet, to find out how the nuances of the meaning of the word anjay in each tweet. From the k-fold cross validation process, the average accuracy result is 50%, the precision result is 59.2%, the recall result is 47.3%, and the f-measure result is 41.2%. The word anjay is found in phrases or nouns 252 tweets, verbs 232 tweets, adjectives 187 tweets, pronouns 32 tweets, numerals 16 tweets, prepositions 12 tweets, and adverbs 11 tweets. The word anjay is found in 350 tweets of declarative sentences, 283 exclamative tweets, 78 interrogative tweets, and 31 imperative tweets. The word anjay has various interjections, including; interjection of admiration, interjection of wonder, interjection of gratitude, interjection of surprise, interjection of hope, interjection of anger, interjection of curses, and interjection of hope. In terms of nuances of meaning using redundancy, the word anjay has no influence in a sentence. The word anjay serves as an affirmation of the core of a sentence.

Optimization Prediction of Big Five Personality in Twitter Users

Various kinds of information can be acquired from social media platforms; one of them is on Twitter. User biographical information and tweets are the essential assets for research that can describe the Big Five Personality, including openness, conscientiousness, extraversion, agreeableness, and neuroticism. Several previous studies have tried the prediction of Big Five Personality. However, the authors found problems in how to optimize the work of the personality prediction system. So, in this study, Big Five Personality predictions were carried out on users of Twitter and improved the performance of the personality prediction system. We implement optimization techniques such as sampling, feature selection, and hyperparameter tuning to enhance the performance. This study also applies linguistic feature extraction, such as LIWC and TF-IDF. By using 287 Twitter users that have permitted their data to be crawled acquired from an online survey using Big Five Inventory (BFI), and applying all optimization techniques, the average accuracy result is 84.22% which is a 74.44% gain over the specified baseline.

Deep learning of resting-state electroencephalogram signals for three-class classification of Alzheimer’s disease, mild cognitive impairment and healthy ageing

Objective. This study aimed to produce a novel deep learning (DL) model for the classification of subjects with Alzheimer’s disease (AD), mild cognitive impairment (MCI) subjects and healthy ageing (HA) subjects using resting-state scalp electroencephalogram (EEG) signals. Approach. The raw EEG data were pre-processed to remove unwanted artefacts and sources of noise. The data were then processed with the continuous wavelet transform, using the Morse mother wavelet, to create time-frequency graphs with a wavelet coefficient scale range of 0–600. The graphs were combined into tiled topographical maps governed by the 10–20 system orientation for scalp electrodes. The application of this processing pipeline was used on a data set of resting-state EEG samples from age-matched groups of 52 AD subjects (82.3 ± 4.7 years of age), 37 MCI subjects (78.4 ± 5.1 years of age) and 52 HA subjects (79.6 ± 6.0 years of age). This resulted in the formation of a data set of 16197 topographical images. This image data set was then split into training, validation and test images and used as input to an AlexNet DL model. This model was comprised of five hidden convolutional layers and optimised for various parameters such as learning rate, learning rate schedule, optimiser, and batch size. Main results. The performance was assessed by a tenfold cross-validation strategy, which produced an average accuracy result of 98.9 ± 0.4% for the three-class classification of AD vs MCI vs HA. The results showed minimal overfitting and bias between classes, further indicating the strength of the model produced. Significance. These results provide significant improvement for this classification task compared to previous studies in this field and suggest that DL could contribute to the diagnosis of AD from EEG recordings.

A Machine Learning Approach for Improving the Performance of Network Intrusion Detection Systems

Intrusion detection systems (IDS) are used in analyzing huge data and diagnose anomaly traffic such as DDoS attack; thus, an efficient traffic classification method is necessary for the IDS. The IDS models attempt to decrease false alarm and increase true alarm rates in order to improve the performance accuracy of the system. To resolve this concern, three machine learning algorithms have been tested and evaluated in this research which are decision jungle (DJ), random forest (RF) and support vector machine (SVM). The main objective is to propose a ML-based network intrusion detection system (ML-based NIDS) model that compares the performance of the three algorithms based on their accuracy and precision of anomaly traffics. The knowledge discovery in databases (KDD) methodology and intrusion detection evaluation dataset (CIC-IDS2017) are used in the testing which both are considered as a benchmark in the evaluation of IDS. The average accuracy results of the SVM is 98.18%, RF is 96.76% and DJ is 96.50% in which the highest accuracy is achieved by the SVM. The average precision results of the SVM is 98.74, RF is 97.96 and DJ is 97.82 in which the SVM got a higher average precision compared with the other two algorithms. The average recall results of the SVM is 95.63, RF is 97.62 and DJ is 95.77 in which the RF achieves the highest average of recall than SVM and DJ. In overall, the SVM algorithm is found to be the best algorithm that can be used to detect an intrusion in the system.

PENENTUAN KESESUAIAN LAHAN PERTANIAN TANAMAN CABAI MENGGUNAKAN METODE NAÏVE BAYES DI KABUPATEN KUPANG

Tanaman cabai merupakan salah satu tanaman yang dibudidayakan di Nusa Tenggara Timur (NTT). Wilayah Kabupaten Kupang merupakan salah satu wilayah penghasil cabai yang ada di NTT. Produksi cabai secara keseluruhan di Kabupaten Kupang pada tahun 2019 sampai 2020 mengalami peningkatan. Namun peningkatan produksi ini belum maksimal mengingat banyaknya lahan yang belum dimanfaatkan sebagai lahan pertanian. Oleh karena itu dibutuhkan sebuah sistem yang membantu dalam menentukan kesesuaian lahan pertanian untuk penanaman cabai. Dalam penelitian ini penulis menerapkan penalaran berbasis kasus/case-based reasoning (CBR) dalam menentukan kesesuaian lahan pertanian tanaman cabai. Metode yang digunakan dalam penelitian ini adalah Naïve Bayes dengan 7 kriteria yaitu: curah hujan, drainase, tekstur tanah, kedalaman tanah, C-organik, kemiringan lahan dan bahaya terjadinya bencana. Proses pencarian solusi dimulai dengan mengeliminasi data yang tidak relevan menggunakan metode Naive Bayes dan berlanjut dengan perankingan nilai kemiripan terbaik menggunakan KNN. Berdasarkan hasil pengujian dengan 110 kasus lahan cabai didapatkan hasil akurasi tertinggi sebesar 92.2% dan rata-rata hasil akurasi dari keseluruhan fold sebesar 89.1%.

Automatic Detection and Classification of Focal Liver Lesions Based on Deep Convolutional Neural Networks: A Preliminary Study

With the increasing daily workload of physicians, computer-aided diagnosis (CAD) systems based on deep learning play an increasingly important role in pattern recognition of diagnostic medical images. In this paper, we propose a framework based on hierarchical convolutional neural networks (CNNs) for automatic detection and classification of focal liver lesions (FLLs) in multi-phasic computed tomography (CT). A total of 616 nodules, composed of three types of malignant lesions (hepatocellular carcinoma, intrahepatic cholangiocarcinoma, and metastasis) and benign lesions (hemangioma, focal nodular hyperplasia, and cyst), were randomly divided into training and test sets at an approximate ratio of 3:1. To evaluate the performance of our model, other commonly adopted CNN models and two physicians were included for comparison. Our model achieved the best results to detect FLLs, with an average test precision of 82.8%, recall of 93.4%, and F1-score of 87.8%. Our model initially classified FLLs into malignant and benign and then classified them into more detailed classes. For the binary and six-class classification, our model achieved average accuracy results of 82.5 and73.4%, respectively, which were better than the other three classification neural networks. Interestingly, the classification performance of the model was placed between a junior physician and a senior physician. Overall, this preliminary study demonstrates that our proposed multi-modality and multi-scale CNN structure can locate and classify FLLs accurately in a limited dataset, and would help inexperienced physicians to reach a diagnosis in clinical practice.

The Design of Neural Network Based Accuracy Prediction for Software Testing Black-Box A Method with the Equivalence Partitions Technique

The absence of a neural network algorithm model to predict the level of accuracy in terms of black-box software testing, equivalence partitions technique is a problem in this research. In this case, the algorithm used for predicting software errors by researchers is the neural network algorithm and testing the software uses the black-box method with the equivalence partitions technique. The neural network algorithm is an artificial neural system, or neural network are the physical cellular system that can acquire, store and use the knowledge gained from experience for activation using bipolar sigmoid output values which range between -1 to 1. Software testing black-box methods is a testing approach where the data comes from defined functional requirements regardless of the final program structure, and the technique used is equivalence partitions. The design prediction accuracy of this research is by determining the college application to be the software to be tested, then tested using the black-box method with the equivalence partitions technique (this test chosen because it will find software errors in several categories, including functions error or missing, interface errors, errors in data structures or external database access, performance errors, initialization errors and terminations), from the black-box test the dataset obtained. This dataset measures the accuracy of the real output and prediction output. The last step is calculating the error, RSME from the real output and prediction output. The results of this research show that the neural network algorithm was being to measure the accuracy level of software testing applied to determine the prediction of the accuracy level of black-box software testing with the equivalence partitions technique, and the average accuracy results are above 80%.

Implementation of Modified Backpropagation with Conjugate Gradient as Microarray Data Classifier with Binary Particle Swarm Optimization as Feature Selection for Cancer Detection

Cancer is one of the deadliest diseases in the world that needs to be handled as early as possible. One of the methods to detect the presence of cancer cells early on is by using microarray data. Microarray data can store human gene expression and use it to classify cancer cells. But one of the challenges of using microarray is its vast number of features, not proportional to its small number of samples. To resolve that problem, dimensionality reduction is needed to reduce the number of features stored in microarray data. Binary Particle Swarm Optimization (BPSO) is one of the methods to reduce dimensionality of microarray data that can increase classification performance. Although when combined with Backpropagation, BPSO still shows a relatively low performance. In this research, Modified Backpropagation with Conjugate Gradient is used to classify data that has been reduced with BPSO. The average accuracy result of BPSO+CGBP is 86.1%, giving it an improvement compared to BPSO+BP which averaged to 80.8%.

Identification and Recognition of Rotavirus- A Particles in Microscopic Images using Enhanced Hybrid Segmentation Methods

The medical image processing or the medical imaging helps in identifying the different types of diseases that are infected to human body. We know that virus is an infective agent that cannot be seen by the human eyes and this can be viewed with help of a microscope. Rotavirus is type of virus that affect human infants that causes diarrhea, which leads to death in severe condition. Identifying the virus particles in the microscope image is a tedious task. Giving prescription based on the medical diagnosis is dependent how accurately it is done. In this paper, the study of identification, segmentation and feature extraction of the Rotavirus-A in the electron microscopic images is considered. The proposed research method hybridizes four segmentation methods which identifies and classify the rotavirus-A particles in the microscopic image based on their features. The classification is carried out using Decision tree classifier and the accuracy rate is measured. The proposed method yields 96% average accuracy result and this can be improved by considering more dataset and training on it.

Fishing Vessels Behavior Identification for Combating IUU Fishing: Enable Traceability at Sea

Overfishing is a critical catastrophe to the ecosystem and the global food chain. The leading causes are Illegal Unreported and Unregulated Fishing (IUU Fishing) linked to illegal labor. EU and the US have set up the fisheries policy that emphasis on traceability. The traceability principle is to monitor the entire seafood supply chain (Sea to Table). FAO’s technology gap analysis reveals that there is a lack of reliable and affordable automated systems or a lack of links to traceability. The challenge of traceability is tracing back to the catch source with existing data and technology. This study aims at the novel concept of a combination of global and local features of trajectory data for fishing vessel behavior identification and enabling seafood transparency. We present a new technique on a local feature of time series and transform the trajectory pattern to global features for Deep Learning. We apply this technique to AIS and VMS data of Thai fishing vessels (Surrounding Nets, Trawl, Longliner, and Reefer). Fishing vessel behaviors were classified as Fishing, Non-fishing, and Transshipment. Our proposed method gives a robust average accuracy result (97.50%). This concept could solve the IUU Fishing and enable traceability at sea, including monitoring, maritime, and marine resources conservation systems.

Development of a low-cost microcomputer based vein imaging system

In this study, a low-cost microcomputer based dorsal hand-vein imaging system was developed by utilizing non-invasive near infrared imaging technology for the patients having hand vein visibility problems. On the microcomputer, we established a new database with high-quality dorsal-hand vein images. Pre- and post-processing were accomplished in Python language using Open-CV library on the microcomputer. Then, the specialist examined the whole obtained raw vein images to detect blood veins by marking the vein areas on a touch screen monitor, namely the reference images. The accuracy of the vein segmentation of our imaging system was assessed by comparing the reference images and the processed images on the microcomputer. We offered a novel simple pixel to pixel algorithm to test the segmented vein areas by the system. In addition, 2-D cross correlation algorithm was used for a secondary analysis. The average accuracy result of the 2-D cross correlation algorithms was 83.84% while pixel-based algorithm offered 94.47%. According to the results, we may conclude that the proposed vein imaging system do not produce suspicious unrelated vein areas which might create another clinical problem for the patients.

Swarm intelligent based online feature selection (OFS) and weighted entropy frequent pattern mining (WEFPM) algorithm for big data analysis

During the past two decades, frequent pattern mining (FPM) has acquired the interests of many researchers: which involves extracting the itemsets from transactions, sequences from big dataset, which occurs frequently and to recognize from the molecular structures, the common sub graph. In this big data era, the unpredictable flow and huge quantity of data brings new challenges in FPM such as space and time complexity. In general, most of the research work focus on recognizing the patterns that occurs frequently, from the set of specific data, where the patterns within every transaction were definitely known a priori. Among these, the users focus only on the small part of this FP. In order to tackle such problems in the current scenario, it is necessary sometimes to select the important features alone, using appropriate FPM algorithms, in order to reduce the complexity level. The major objective of this work is to improve FPM mining results and improve classification accuracy of big dataset samples. To tackle the first challenge, the levy flight bat algorithm (LFBA) along with online feature selection (OFS) approach is proposed, which is used to filter the low quality features from the big data in an online manner. Subsequently to address the second challenge, a weighted entropy frequent pattern mining (WEFPM) is enforced for FPM, to accomplish better computation time when compared with other methods such as direct discriminative pattern mining (DDPMine) and iterative sampling based frequent itemset mining (ISbFIM), where enumeration of entire feature combinations were completed. So the WEFPM algorithm employed in this paper, targets to identify only the specific frequent patterns which are required by the user. By iterating this procedure, it assures that the acquired frequent patterns can be enumerated by using both the theoretical and empirical research, so that enumeration doesn’t proceed into a combinatorial explosion. And also, using the above said LFBA–OFS approach and WEFPM algorithm, frequent patterns that are different in nature, are generated for building high quality learning model. For finding the frequent patterns, here the minimum support threshold is matched with entropy. As a final step, multiple Kernel learning support vector machine is employed as a classifier, to evaluate the performance of the big data samples for efficiency and accuracy. Empirical study reveal that considerable progress is obtained in terms of accuracy and computation time when applied to UCI benchmark big datasets, using the proposed approach for efficient and effective FPM of the online features. It is clear that WEFPM is the most efficient method, because it produces higher average accuracy results of 92.34, 93.218, 91.374 and 87.87% values for adult, chess, hybo and sick dataset respectively. It outperforms when compared to other methods such as DDPMine and ISbFIM using an LIBSVM classifier.

Real-Time Electroencephalography-Based Emotion Recognition System

This paper proposes parametric, general and effectively automatic real time classification method of electroencephalography (EEG) signals based on emotions. The specific characteristics of the high-frequency signals (alpha, beta, gamma) are observed, and then Fourier Transform, Features Extraction (mean, standard deviation, power) and the K-Nearest Neighbors (KNN) are employed for signal processing, analysis and classification. The proposed method consists of two stages for a multi-class classification and it can be considered as the framework of multi-emotions based on Brain Computer Interface (BCI). The first stage, the calibration, is off-line and it computes the signal processing, determines the features and trains the classification. The second stage, the real-time, is the test on new data. The FFT is applied to avoid redundancy in the selected features; then the classification is carried out using the KNN. The results show that the average accuracy results are 82.33% (valence) and 87.32% (arousal).

MODELLING OF CARBON MONOXIDE AIR POLLUTION IN LARG CITIES BY EVALUETION OF SPECTRAL LANDSAT8 IMAGES

Abstract. Air pollution in large cities is one of the major problems that resolve and reduce it need multiple applications and environmental management. Of The main sources of this pollution is industrial activities, urban and transport that enter large amounts of contaminants into the air and reduces its quality. With Variety of pollutants and high volume manufacturing, local distribution of manufacturing centers, Testing and measuring emissions is difficult. Substances such as carbon monoxide, sulfur dioxide, and unburned hydrocarbons and lead compounds are substances that cause air pollution and carbon monoxide is most important. Today, data exchange systems, processing, analysis and modeling is of important pillars of management system and air quality control. In this study, using the spectral signature of carbon monoxide gas as the most efficient gas pollution LANDSAT8 images in order that have better spatial resolution than appropriate spectral bands and weather meters،SAM classification algorithm and Geographic Information System (GIS ), spatial distribution of carbon monoxide gas in Tehran over a period of one year from the beginning of 2014 until the beginning of 2015 at 11 map have modeled and then to the model valuation ،created maps were compared with the map provided by the Tehran quality comparison air company. Compare involved plans did with the error matrix and results in 4 types of care; overall, producer, user and kappa coefficient was investigated. Results of average accuracy were about than 80%, which indicates the fit method and data used for modeling.

Initialism disambiguation: Man versus machine

Disambiguation of ambiguous initialisms and acronyms is critical to the proper understanding of various types of texts. A model that attempts to solve this has previously been presented. This model contained various baseline features, including contextual relationship features, statistical features, and language‐specific features. The domain of Jewish law documents written in Hebrew and Aramaic is known to be rich in ambiguous abbreviations and therefore this model was implemented and applied over 2 separate corpuses within this domain. Several common machine‐learning (ML) methods were tested with the intent of finding a successful integration of the baseline feature variants. When the features were evaluated individually, the best averaged results were achieved by a library for support vector machines (LIBSVM); 98.07% of the ambiguous abbreviations, which were researched in the domain, were disambiguated correctly. When all the features were evaluated together, the J48 ML method achieved the best result, with 96.95% accuracy. In this paper, we examine the system's degree of success and the degree of its professionalism by conducting a comparison between this system's results and the results achieved by 39 participants, highly fluent in the research domain. Despite the fact that all the participants had backgrounds in religious scriptures and continue to study these texts, the system's accuracy rate, 98.07%, was significantly higher than the average accuracy result of the participants, 91.65%. Further analysis of the results for each corpus implies that participants overcomplicate the required task, as well as exclude vital information needed to properly examine the context of a given initialism.