Average Accuracy Rate Research Articles

New Arabic Root Extraction Algorithm

This research presents a new algorithm for Arabic root extraction, which aims to improve the accuracy of Arabic Natural Language Processing Algorithms by addressing the weaknesses and errors of existing algorithms. The proposed algorithm utilizes a database, that includes a collection of roots, patterns, and affixes, to generate potential derivation roots of a word without eliminating affixes initially. By matching the derived word with patterns to identify potential roots, the proposed algorithm avoids the inaccuracies caused by eliminating affixes based on expectation methods. The study conducted a comparison of the proposed algorithm with three commonly used Arabic root extraction algorithms. The evaluation process is performed on three corpora. Results showed that the proposed algorithm achieved an average accuracy rate of 96%, which is significantly higher than the others.

The Separation of glaucoma and non-glaucoma fundus images using EfficientNet-B0

Glaucoma is an eye disease that causes vision loss. This disease progresses silently without symptoms. Therefore, it is a difficult disease to detect. If glaucoma is detected before it progresses to advanced stages, vision loss can be prevented. Computer-aided diagnosis systems are preferred to understand whether the fundus image contains glaucoma. These systems provide accurate classification of healthy and glaucoma images. In this article, a system to separate images of a fundus dataset as glaucoma or healthy is proposed. The EfficientNet B0 model, which is a deep learning model, is used in the proposed system. The input of this deep network model is designed as six layers. The experimental results of the designed model were obtained on the publicly available ACRIMA dataset images. In the end, the average accuracy rate is determined as 0.9775.

Covid-19 Kısıtlamaları Sırasında Duyguların Müzikal Parametrelerle İfade Edilmesi: Filipinler Spotify Verileri Üzerine Bir Duygu Analizi

Spotify is the world's largest online music streaming platform that offers a tremendous variety of playlists based on listeners' listening patterns. This paper proposes that music preference is highly associated with emotional state, and music is an emotion regulator tool during the pandemic in the Philippines. Well-known machine learning methods (i.e., classification and regression trees, boosted trees, random forests, Support Vector Machines, and Artificial Neural Networks) in combination with 5-fold cross-validation are used to classify periods in proportion to the severity of the pandemic and people's musical preferences. Daily official covid-19 statistics and Spotify data are used as main variables during the algorithms' learning processes. SVM outperformed the other alternatives in average accuracy rate by achieving a 98.01% accuracy rate. Additionally, ANN outperformed the other alternatives in terms of accuracy achieved specifically in a single model, achieving an accuracy rate of 99.30%. Moreover, the variables with the largest (absolute) change (in descending order) are ST_Intrumentalness (26,45%), ST_Acousticness (19,03%), ST_Liveness (16,11%), and ST_Valence (14,1%). Given pandemics-related stress and cancelation of concerts, it would be an intuitive expectation that the variables ST_Valence (musical positivity) and ST_Liveness would change at such a rate. The results confirm that musical preference is a significant indicator of emotional state.

基于主成分分析的猪脸特征提取研究

Individual identification and behavioural analysis of pigs is a key link in the intelligent management of a piggery, for which the computer vision technology based on application and improvement of deep learning model has become the mainstream. However, the operation of the model has high requirements to hardwares, also the model is of weak interpretability, which make it difficult to adapt to both the mobile terminals and the embedded applications. In this study, it is first put forward that the key facial features of pigs can be extracted by PCA method first before the eigen face method is adopted for verification tests to reach an average accuracy rate of 74.4%; the key features, for which the most identifiable ones are in turn, respectively, face contour, nose, ears and other parts of pigs, can be visualized, and this is different from the identification features adopted in manual identification. This method not only reduces the computational complexity but also is of strong interpretability, so it is suitable for both the mobile terminals and the embedded applications. In some way, this study provides a systematic and stable guidance for livestock and poultry production.

DESIGN A MONITORING SYSTEM OF A SINGLE-PHASE ELECTRICITY METER BASED ON THE INTERNET OF THINGS

This research aims to create systematic monitoring of a single-phase electricity meter that is accompanied by power monitoring features, the amount of electrical energy usage costs and relay control features using internet connections. The design has been carried out using PZEM-004T current and voltage sensors. The microcontrollers have been implemented by using ESP8266. This configuration is to reduce separate micro-controller. The embedded microcontroller and WiFi are made improvements from previous research. This embedded microcontroller is used to send data to the internet through WIFI. This monitoring system can then be accessed using the Cayenne interface. The test results of this tool show that the connection between the microcontroller and the Cayenne application works well through an internet connection. With this tool, users can see real-time data on the use of electrical energy and its cost, as shown in the cayenne dashboard. The voltage sensor test results show an average accuracy is 99.17%. The current sensor testing has an average accuracy rate of 96.9%. On average, the wattage delta between the cayenne dashboard and multimeter measurement is 2.16 watts.

청각장애인을 위한 인공지능 헤드폰 설계

We designed an artificial intelligence (AI) headphone for hearing impaired people and proposed a convolution neural network (CNN)-based sound classifier with a low computational network that can run on an embedded PC (Raspberry Pi 4B) in real-time. Because our AI headphone can classify 20 types of dangerous or environmental sounds (e.g., siren, car horn, scream, gunshot, etc.) and recognize specific voice keywords (e.g., person name, be careful!, stop!, etc.), it can assist hearing impaired people in reducing the risk of accident exposure and improving convenience. In addition, we use vibration codes to notify the hearing impaired person of detected information to the vibration motors attached to both sides of the headphone. We confirm that our proposed sound classifier achieves an average accuracy rate of approximately 95.14%, and enables real-time processing on the Raspberry Pi 4B because it requires an average computation time of approximately 0.139s with audio recording data for 5.12s.

Fire Detection on Video Using ViBe Algorithm and LBP-TOP

In this research, we built a system to detect fire using the ViBe (Visual Background Extractor) algorithm to extract dynamic targets. The ViBe algorithm is better at detecting moving target objects such as flame combustion. In this research we combined the ViBe algorithm with three frame differencing to gain better results on movement object. The HSI color space model was applied after the movement object was obtained. We used Local Binary Pattern-Three Orthogonal Planes to obtain the feature extraction to be classified with Support Vector Machine. Our result has shown that the proposed system were able to detect the fire using the LBP-TOP and ViBe algorithm methods with an average accuracy rate of 88.10%, and the best accuracy was 90.37%. The parameters used to achieve this accuracy in the feature extraction process were T=120, Radius=2, and frame gap=15, then the threshold value parameter for three-frame difference parameter was 25.

Guidelines to build PLS-DA chemometric classification models using a GC-IMS method: Dry-cured ham as a case of study

The number of representative samples to build a calibration model plays a major role in the success of chemometric models for class discrimination; therefore, knowing which samples should be used for the calibration of prediction models is essential. The aim of this work is to design a basic guideline for the training of partial least squares discriminant analysis (PLS-DA) models to classify complex samples analysed by Gas Chromatography (GC) coupled to Ion Mobility Spectrometry (IMS) using dry-cured Iberian ham as an example. The effect of the number, proportion and class of samples for training and validation and the use of two data types (spectral fingerprint or pre-selected markers) has been assessed by analysing with GC-IMS nearly 1000 dry-cured Iberian ham samples obtained from 7 different curing plants. Subsequently, these were classified with PLS-DA according to the pig's feeding regime (acorn-fed vs. feed-fed) and it has been demonstrated that 450 out of 997 samples are enough for model training to achieve a maximum average prediction accuracy rate. Furthermore, the use of pre-selected GC-IMS markers provides slightly better prediction results than the use of the complete spectral fingerprint. In summary, these results represent a tentative guide for the classification of samples in an industrial setting using GC-IMS and PLS-DA. This methodology would allow authorities and producers to ensure the quality of the agri-food products put on the market as is proven in this study.

Fault diagnosis of bearing based on refined piecewise composite multivariate multiscale fuzzy entropy

As one of the key components of the train, the condition of the bearing is related to the train's safe operation. The vibration signal of the bearing is usually nonlinear and nonstationary, which makes it difficult to extract fault features and leads to low diagnostic accuracy. Entropy theory can effectively measure the change of nonlinearity and complexity of the vibration signal. However, the conventional entropy algorithm cannot extract the characteristics of a multi-channel signal simultaneously. A bearing fault diagnosis method based on refined piecewise composite multivariate multiscale fuzzy entropy (RPCMMFE) and convolutional neural network (CNN) is proposed. The proposed method can fully extract fault features and improve fault diagnosis accuracy. Firstly, based on multivariate multiscale fuzzy entropy (MMFE), RPCMMFE is proposed by introducing refined theory and piecewise composite theory. Then, the RPCMMFE of different fault signals is calculated, and RPCMMFE is considered to be a feature vector that acts as an input into the CNN for fault diagnosis. Finally, it is verified by two groups of experiments. The experimental results show that the features extracted by this method have better stability and discrimination ability and can identify bearing faults more accurately. The average accuracy rates are 99% and 99.17%, respectively.

Intrusion Detection in Big Data Era: A Multi- Objective Approach for Longer Model Lifespans

Despite highly accurate intrusion detection schemes based on machine learning (ML) reported in the literature, changes in network traffic behaviour quickly yield low accuracy rates. An intrusion detection model update is not easily feasible due to the enormous amount of network traffic to be processed in near real-time for high- speed networks, in particular, under big data settings. In this paper, we propose a new scalable long-lasting intrusion detection architecture for the processing of network content and the building of a reliable ML-based intrusion detection model. Experiments performed through the analysis of five years of network traffic, about 20 TB of data, have shown that our approach extends the lifespan of our model by up to six weeks. That occurs because the average accuracy rate of our proposal lasted eight weeks after the training phase, and traditional ones reached only two weeks after the model building. Additionally, our proposal achieves up to 10 Gbps of detection throughput in a 20-core big data processing cluster. KEYWORDS: Machine Learning, Intrusion Detection, Big Data, High-Speed Networks.

Hybrid Deep Model for Automated Detection of Tomato Leaf Diseases

Tomatoes are preferred by farmers because of their high productivity. This fruit has a fibrous structure and contains plenty of vitamins. Tomato diseases are generally observed on stem, fruit, and leaves. Early diagnosis of the disease in plants is of vital importance for the plant. This is very important for farmers who expect economic gain from that plant. Because if the disease is not treated early, these tomatoes should be destroyed. For these reasons, systems to diagnose the disease early are very important. In this study, a tomato leaf diseases classification model developed with deep learning methods, which is one of the most popular artificial intelligence techniques, is proposed in order to eliminate the possibility of the human eye being mistaken. In this study, 6 different Convolutional Neural Network (CNN) architectures were used. In the first stage of this study, which consists of two stages, the classification process was carried out with the Alexnet, Googlenet, Shufflenet, Efficientb0, Resnet50, and Inceptionv3 architectures that were previously trained. In the second stage, feature maps of tomato leaf images in the dataset were obtained using the six pre-trained deep learning architectures. In the hybrid model proposed in this study, the feature maps extracted using the best two of the six deep learning models are concatenated. Then, the Neighborhood Component Analysis (NCA) method was applied to the extracted features in order to speed up the system, unnecessary features were removed and optimized. The optimized feature map is classified by traditional intelligent classification models. As a result of experimental studies, the average accuracy rate of the proposed model is 99.50 percent.

Teachable Machine: Real-Time Attendance of Students Based on Open Source System

The utilization of open source-based services will be very useful, simplifying and accelerating the process of object recognition and complex computational processes, one of them uses the Teachable Machine service. Identification of student faces in real-time attendance is a case study that will be applied to students to recognize and identify accurately and clearly the presence of students during online / offline lectures, by applying Teachable Machine services that have good algorithms with a machine learning approach that utilizes the Tensorflow.js library where the training data testing uses Convolutional Neural Network (CNN). Of the objects identified, the average accuracy of all classes ranged from 91-100%, with the number of samples for each object class being 23 objects or more. Number of sample images in one class. Clothing, object background and lighting intensity around the image object are also very influential in determining the accuracy value of student face recognition later, so that the use of the tensorflow.js library that implements Convolutional Neural Network (CNN) will be very helpful in facial recognition and influencing factors so that the data entered later needs to be further corrected and improved again, so that the results obtained in implementing the online attendance system have been very helpful in detecting student faces with an average accuracy rate of 91.8%

Optimization of Hyperparameters in Object Detection Models Based on Fractal Loss Function

Hyperparameters involved in neural networks (NNs) have a significant impact on the accuracy of model predictions. However, the values of the hyperparameters need to be manually preset, and finding the best hyperparameters has always puzzled researchers. In order to improve the accuracy and speed of target recognition by a neural network, an improved genetic algorithm is proposed to optimize the hyperparameters of the network by taking the loss function as the research object. Firstly, the role of all loss functions in object detection is analyzed, and a mathematical model is established according to the relationship between loss functions and hyperparameters. Secondly, an improved genetic algorithm is proposed, and the feasibility of the improved algorithm is verified by using complex fractal function and fractional calculus. Finally, the improved genetic algorithm is used to optimize the hyperparameters of the neural network, and the prediction accuracy of the model before and after the improvement is comprehensively analyzed. By comparing with state-of-the-art object detectors, our proposed method achieves the highest prediction accuracy in object detection. Based on an average accuracy rate of 95%, the detection speed is 20 frames per second, which shows the rationality and feasibility of the optimized model.

A Novel Method for Network Intrusion Detection

Intrusion detection is one of the key research directions of network information security. In order to make up for the deficiencies of traditional security technologies such as firewall, encryption, and authentication, by analyzing the characteristics of network attacks and existing intrusion detection models, the advantages of triadic concept analysis and the application of fuzzy set theory in network intrusion detection are analyzed. The intrusion detection model FCTA based on triadic concept analysis is proposed, which promotes the further development of network intrusion detection. First, we analyze the characteristics of the data and use TF-IDF and Z-Score to normalize and standardize the data to construct a fuzzy triadic background containing quadratic characteristics. It is used to describe the triadic relationship between network connections, network connection characteristics, and intrusion types of network packets. Then, the (i)-induced operator is used to construct the fuzzy triadic concept set based on the fuzzy triadic background and transformed into a fuzzy attribute triadic concept set. Then, the new samples are classified by calculating the similarity between the new samples and the elements in the fuzzy attribute triadic concept set by using the Euclidean distance formula. In order to reduce the model space complexity, compression storage technology is adopted in the model building process.. Finally, by using the IDS-2018 dataset, the rationality and effectiveness of the FCTA model are demonstrated. The average accuracy and average intrusion detection rate of FCTA classification are much higher than BP neural network, SVM algorithm, and KNN algorithm, and the FCTA misjudgment rate is much lower than the BP neural network algorithm, the KNN algorithm, and the SVM algorithm; with the increase of data volume, the accuracy rate and intrusion detection rate increase significantly.

GA-BP neural network modeling for project portfolio risk prediction

PurposeProject portfolio risk (PPR) management plays an important role in promoting the smooth implementation of a project portfolio (PP). Accurate PPR prediction helps managers cope with risks timely in complicated PP environments. However, studies on accurate PPR impact degree prediction, which consists of both risk occurrence probabilities and risk impact consequences considering project interactions, are limited. This study aims to model PPR prediction and expand PPR prediction tools.Design/methodology/approachIn this study, the authors build a PPR prediction model based on a genetic algorithm and back-propagation neural network (GA-BPNN) integrated with entropy-trapezoidal fuzzy numbers. Then, the authors verify the proposed model with real data and obtain PPR impact degrees.FindingsThe test results indicate that the proposed method achieves an average absolute error of 0.002 and an average prediction accuracy rate of 97.8%. The former is reduced by 0.038, while the latter is improved by 32.1% when compared with the results of the original BPNN model. Finally, the authors conduct an index sensitivity analysis for identifying critical risks to effectively control them.Originality/valueThis study develops a hybrid PPR prediction model that integrates a GA-BPNN with entropy-trapezoidal fuzzy numbers. The authors use this model to predict PPR impact degrees, which consist of both risk occurrence probabilities and risk impact consequences considering project interactions. The results provide insights into PPR management.

Model Simulasi untuk Memperkirakan Tingkat Penjualan Garam Menggunakan Metode Monte Carlo

The increasing need for salt in the West Sumatra area is inversely proportional to the raw material for making salt. So the stock of salt for consumption becomes less. This causes the purchasing power to be cut off for the needs of regional consumers. Based on these problems, a research was conducted by conducting simulations to predict the amount of salt sales in controlling stock. This study aims to predict sales in maintaining service to consumer demand. The method that can be used in making predictions is the Monte Carlo Method by processing Salt sales data in 2019, 2020, and 2021 at PT. Prosperous Grace. The results of the study are able to predict sales of salt in the form of kilograms (kg) in the future. The average accuracy rate in 2020 is 88% and in 2021 is 91%. So that this research can be a reference in decision making by PT. Kurnia Sejahtera to improve services.

Real-time object detection and selection with the LabVIEW program

Nowadays, the demand for production is increasing due to the increase in the human population. For this reason, different developing technologies are used to meet the required production. In developing technologies, image processing techniques are used to save manpower and time and to minimize possible errors. In this study, image processing techniques were used to detect and select the colors and shapes of the objects coming over the conveyor belt system. Real-time images were preferred in the study. In the implemented system, the selection process was carried out by using the LabVIEW program to define the colors and shapes of the objects. LabVIEW NI- IMAQdx was used to find the colors of the objects. To facilitate the definition of shapes, the images taken from the Vision Assistant module in the LabVIEW program were converted to HSL format, and shape definitions were made using different algorithms. After these processes were done, the servo motors in the conveyor belt system were enabled to communicate with each other with the help of the Arduino program and the selection of the objects was carried out. According to the results obtained, the average accuracy rate for three-dimensional objects was 95.349 %. This rate is considered to be a very high rate for object detection of correct color and shape.

New Classification Method for Independent Data Sources Using Pawlak Conflict Model and Decision Trees.

The research concerns data collected in independent sets-more specifically, in local decision tables. A possible approach to managing these data is to build local classifiers based on each table individually. In the literature, many approaches toward combining the final prediction results of independent classifiers can be found, but insufficient efforts have been made on the study of tables' cooperation and coalitions' formation. The importance of such an approach was expected on two levels. First, the impact on the quality of classification-the ability to build combined classifiers for coalitions of tables should allow for the learning of more generalized concepts. In turn, this should have an impact on the quality of classification of new objects. Second, combining tables into coalitions will result in reduced computational complexity-a reduced number of classifiers will be built. The paper proposes a new method for creating coalitions of local tables and generating an aggregated classifier for each coalition. Coalitions are generated by determining certain characteristics of attribute values occurring in local tables and applying the Pawlak conflict analysis model. In the study, the classification and regression trees with Gini index are built based on the aggregated table for one coalition. The system bears a hierarchical structure, as in the next stage the decisions generated by the classifiers for coalitions are aggregated using majority voting. The classification quality of the proposed system was compared with an approach that does not use local data cooperation and coalition creation. The structure of the system is parallel and decision trees are built independently for local tables. In the paper, it was shown that the proposed approach provides a significant improvement in classification quality and execution time. The Wilcoxon test confirmed that differences in accuracy rate of the results obtained for the proposed method and results obtained without coalitions are significant, with a p level = 0.005. The average accuracy rate values obtained for the proposed approach and the approach without coalitions are, respectively: 0.847 and 0.812; so the difference is quite large. Moreover, the algorithm implementing the proposed approach performed up to 21-times faster than the algorithm implementing the approach without using coalitions.

Kernel Density Estimation and Convolutional Neural Networks for the Recognition of Multi-Font Numbered Musical Notation

Optical music recognition (OMR) refers to converting musical scores into digitized information using electronics. In recent years, few types of OMR research have involved numbered musical notation (NMN). The existing NMN recognition algorithm is difficult to deal with because the numbered notation font is changing. In this paper, we made a multi-font NMN dataset. Using the presented dataset, we use kernel density estimation with proposed bar line criteria to measure the relative height of symbols, and an accurate separation of melody lines and lyrics lines in musical notation is achieved. Furthermore, we develop a structurally improved convolutional neural network (CNN) to classify the symbols in melody lines. The proposed neural network performs hierarchical processing of melody lines according to the symbol arrangement rules of NMN and contains three parallel small CNNs called Arcnet, Notenet and Linenet. Each of them adds a spatial pyramid pooling layer to adapt to the diversity of symbol sizes and styles. The experimental results show that our algorithm can accurately detect melody lines. Taking the average accuracy rate of identifying various symbols as the recognition rate, the improved neural networks reach a recognition rate of 95.5%, which is 8.5% higher than the traditional convolutional neural networks. Through audio comparison and evaluation experiments, we find that the generated audio maintains a high similarity to the original audio of the NMN.

다수 조명의 채널별 융합을 이용한 CNN 기반 머신 비전 분류기

In this paper, a method of combining three grayscale images into a color image and applying it to a recent CNN-based classifier is proposed. Each of the three color channels is assigned by three gray-scale images of a target object taken under different lighting conditions using the same optical system. Consequently, a single virtual color image is obtained. By creating a virtual color image of the classified ball, the generated images and the original gray-scale image used for the machine vision inspection are composed as a dataset, and the classification results are compared by learning with ResNet-50. In the case of bump ball pressing/protrusion, the accuracy of defect classification is computed by conducting experiments five times. Through a comparative evaluation, the accuracy of general grayscale image classification was determined as between 80.2% and 95.5% (depending on lighting conditions), and the average accuracy rate was 89.6%. According to the combination method, the accuracy rate of the virtual color image was distributed between 94.2% and 97.3% and the average accuracy rate was 96.1%. Accordingly, we can confirm a comparative advantage and an average improvement effect of 6.5% in terms of stability and accuracy.