Figure 1- - uploaded by Junita Mohamad-Saleh
Content may be subject to copyright.
Source publication
Most of the real-world data samples used to train artificial neural networks (ANNs) consist of correlated information caused by overlapping input instances. Correlation in sampled data normally creates confusion over ANNs during the learning process and thus, degrades their generalization capability. This paper proposes the Principal Component Anal...
Contexts in source publication
Context 2
... learning process employs a learning algorithm, during which the MLP develops a mapping function between the inputs and outputs. Basically, in a learning process, the input PEs receive data from the external environment (denoted by x 1 , x 2 , … x n in Figure 1) and pass them to the hidden PEs, which are responsible for simple, yet, useful mathematical computations involving the weight of the links (denoted by w 11 , w 21 , … in the figure) and the input values. The results from the hidden PEs are mapped onto appropriate threshold function of each PE and the final outputs are produced. ...
Context 3
... results from the hidden PEs are mapped onto appropriate threshold function of each PE and the final outputs are produced. The output values then become inputs to all PEs in the adjacent layer (either the second hidden layer or the output layer), and the computation processes are repeated through out the layers until finally, output values are produced at the output PEs (denoted by y 1 , y 2 , … in Figure 1). At this stage, an output error value is calculated by computing the difference between the MLP's and the actual outputs. ...
Similar publications
This paper describes the student’s facial emotion recognition to determine their current state of mind. The facial emotion recognition of the student could be envisioned to sense their attention states through a video camera. Therefore, facial images were analyzed to extract the features using Principal Component Analysis (PCA) algorithm. Then, the...
Exploratory Factor Analysis (EFA) and Principal Component Analysis (PCA) are popular techniques for simplifying the presentation of, and investigating the structure of, an (n×p) data matrix. However, these fundamentally different techniques are frequently confused, and the differences between them are obscured, because they give similar results in...
This paper describes the facial emotion recognition to determine the student's conditions. The facial emotion recognition could be envisioned to sense the student's attention state through a CCD camera. Therefore, facial images were analyzed to extract the features to characterize the variations between two categories of student's faces (understand...
Citations
... On the other hand, RNN systems can have this sort of input structure. They come about of a hub can be utilized once more as inputs for prior layers [11][12][13][14][15][16][17][18]. This uncommon highlight may appear up clearly when altering utilizing boisterous estimations, which we are going conversation around afterward [16][17][18][19][20][21][22][23][24][25]. ...
Gastric cancer is an important health problem and is the fourth most common cancer and the second leading cause of cancer-related deaths worldwide. The incidence of stomach cancer is increasing and it can be dealt with using new methods in prediction and diagnosis. Our goal is to implement an artificial neural network to predict new cancer cases. Gastric cancer is anatomically divided into true gastric adenocarcinomas (non-cardiac gastric cancers) and gastric-esophageal- connective cancer (adenocardia (cardiac) gastric cancers). We use MATLAB R2018 software (MathWorks) to implement an artificial neural network. We used. The data were repeatedly and randomly divided into training (70%) and validation (30%) subsets. Our predictions emphasize the need for detailed studies on the risk factors associated with gastric cell carcinoma to reduce the incidence and has also provided an accuracy of about 99.998%.
... Authors [31], the gas classification neural network was trained using 1175 train and 375 test vectors obtained from the SPCA transformed data of their eight element SnO 2 sensor response. As a result, to be able to use PCA/SPCA trained neural classifiers in order to test freshly generated samples, [32], a multi-layer perceptron (MLP) classifier has been developed along with a pre-processor, a postprocessor, and a training heuristic. There are various other application where we can apply this technique [33][34][35]. ...
Gas sensor arrays have been used to identify gases based on their dynamic response. Pre-processing of data plays a critical role in identifying different gases. There are several pre-processing techniques for data found in the literature, all of which are complex and sophisticated. Six different volatile organic compounds were exposed to eight metal-oxide (MOX) gas sensors under tightly controlled operating conditions at different concentration levels. In addition, the generated dataset can be used to tackle a variety of chemical sensing challenges such as gas identification, sensor failure, or system calibration. It is a novel way to classify mixed gases. According to the statistical parameters calculated, the proposed GMDH neural network was found to exhibit acceptable accuracy in this study. The mean square (\({R}^{2}\)) value for the proposed GMDH model obtained (\({R}^{2}\)) train set = 0.99346, (\({R}^{2}\)) test set = 0.992, (\({R}^{2}\)) complete set: 0.99261, represents a good match between the actual and output amounts of the proposed model, which is interpreted as the efficiency of the GMDH algorithm in predicting ethylene in mixtures. It has been verified that the proposed GMDH technique produces accurate classification outcomes.KeywordsClassificationGas sensorMixed gas dataNeural networkRegression
... The use of neural networks in combination with other machine learning techniques (Linear Discriminant Analysis, Principal Component Analysis) improved the NN performance [31][32][33]. The methodology, the main results, and discussion are presented in the following sections of our paper. ...
From an economic point of view, shortly digitalization will lead not only to progressive growth, but also to important transformations of jobs and to the reorganization of the way of carrying out the activity of retail, transport, and banking services. Our research aimed to identify and analyze the correlations between digitization and economic growth of EU countries in the period 2019–2021. In this paper, we use deep learning and principal component analysis as an efficient technique to improve the accuracy of classification for the set of EU countries classified according to The Digital Economy and Society Index. The used databases were Eurostat and World Bank. We selected 15 indicators on which we first trained a 2-layer neural network and we obtained a classifier with 92.52% accuracy. Then, we applied principal component analysis and reduced the original dataset to 3 principal components which retain together 78.21% of the initial variability. We train a 2-layer neural network on the score matrix given by the three retained principal components. The results revealed that the classification improved from 92.52 to 100%.KeywordsDigitalizationEconomic growthPrincipal component analysisNeural networksDeep learning
... By having fewer domain values models can be trained and can run faster. This is seen in research like [5] and [6]. But there are examples of PCA improving accuracy for Neural Networks. ...
This research uses Principal Component Analysis (PCA) in conjunction with a Neural Network to increase the accuracy of predicting student test scores. Much research has been conducted attempting to predict student test scores using a standard, well-known dataset. The dataset includes student demographic and educational data and test scores for Mathematics and Language. Multiple predictive algorithms have been used with a Neural Network being the most common.
In this research PCA was used to reduce the domain space size using varying sizes. This began with just 1 attribute and increased to the full size of the original set’s domain values. The reduced domain values and the original domain values were independently used to train a Neural Network and the Mean Absolute errors
were compared. Because results may vary depending upon which records in the dataset are training versus testing, 50 trials were conducted for each reduction size. Results were average and statistical tests were applied. Results show that using PCA prior to training the Neural Network can decrease the mean absolute error by up to 15%.
... The improvement of the model quality by using a PCA has already been demonstrated in other studies (see, for example, [54]). By performing PCA, the 379 features were transformed into 46 principal components, which explain most of the variance of the original features. ...
Designing customized products for customer needs is a key characteristic of machine and plant manufacturers. Their manufacturing process typically consists of a design phase followed by planning and executing a production process of components required in the subsequent assembly. Production delays can lead to a delayed start of the assembly. Predicting potentially delayed components—we call those components assembly start delayers—in early phases of the manufacturing process can support an on-time assembly. In recent research, prediction models typically include information about the orders, workstations, and the status of the manufacturing system, but information about the design of the component is not used. Since the components of machine and plant manufacturers are designed specifically for the customer needs, we assumed that material data influence the quality of a model predicting assembly start delayers. To analyze our hypothesis, we followed the established CRISP-DM method to set up 12 prediction models at an exemplary chosen machine and plant manufacturer utilizing a binary classification approach. These 12 models differentiated in the utilization of material data—including or excluding material data—and in the utilized machine learning algorithm—six algorithms per data case. Evaluating the different models revealed a positive impact of the material data on the model quality. With the achieved results, our study validates the benefit of using material data in models predicting assembly start delayers. Thus, we identified that considering data sources, which are commonly not used in prediction models, such as material data, increases the model quality.
... Therefore, PCA not only increases the computation speed but also decreases the chance of including redundant information in the network. The PCA-ANN has been applied in other studies (Charytoniuk and Chen 2000;Mohamad-Saleh and Hoyle 2008;Neumann 2002;Oliveira-Esquerre et al. 2002;Tabe et al. 1999) to reduce the dimension of the input space while lowering the training time and preserving or even improving the ANN model accuracy (Zekić-Sušac et al. 2013). ...
Prediction of plantation productivity from environmental data is challenged by the complex relationships defining the growth and yield processes. Artificial neural network (ANN) can be used to model complex nonlinear interrelations but are challenged by the choice of the training algorithm and the structure and size of the network. The objective of the present study is to find an efficient ANN to estimate the eucalyptus productivity from biotic, abiotic and silvicultural data. We defined the efficiency of an ANN as the processing time to supply an accurate solution. We increase the efficiency of the network in two steps: one outside the ANN, through principal component analysis (PCA), and one inside the ANN, through dedicated pruning methods. To test different network configurations, we used data from 507 Eucalyptus plantations, 3 to 7 years old, located in Minas Gerais, Brazil. An ANN configuration is a combination of the number of neurons in the hidden layers, training algorithm and pruning method. Each ANN was trained five times to assess the impact of the initial weights on the results, which led to a factorial experiment with 9000 combinations. The most accurate result was supplied in 38.81 s by an ANN using the data trimmed with PCA, trained with the Scaled Conjugate Gradient algorithm with four neurons and Magnitude-Based Pruning method. However, an accuracy loss of less than 1% (i.e., the second most accurate result) was obtained in 1.7 s using the same ANN configuration, but no pruning. Our results indicate that an efficient prediction of Eucalyptus productivity does not use all the data or the most complex training algorithms. The proposed efficient ANN supplies the most precise results compared with the existing models. Furthermore, the efficient ANN produces more accurate results from environmental variables than some models that contain dimensional variables. Our study suggests that modeling the yield of Eucalyptus plantations should be executed using a large set of environmental variables, as no clear edaphic or climatic variables supplied accurate results.
... Переменные, оказывающие наименьшее влияние на разброс данных, это те, которые могут быть удалены из анализа в попытке уменьшить количество независимых переменных. Примеры эффективного применения анализа главных компонентов для повышения производительности инструментов машинного обучения можно легко найти, например, в работах [14][15][16]. Однако стоит иметь в виду, что PCA не принимает во внимание значение зависимой переменной. Следовательно, нет уверенности в том, что именно независимая переменная, которая также оказывает наибольшее влияние на прогнозируемое значение (зависимая переменная), вызывает наибольшую изменчивость в наборах данных. ...
The use of data mining and machine learning tools is becoming increasingly common. Their usefulness is mainly noticeable in the case of large datasets, when information to be found or new relationships are extracted from information noise. The development of these tools means that datasets with much fewer records are being explored, usually associated with specific phenomena. This specificity most often causes the impossibility of increasing the number of cases, and that can facilitate the search for dependences in the phenomena under study. The paper discusses the features of applying the selected tools to a small set of data. Attempts have been made to present methods of data preparation, methods for calculating the performance of tools, taking into account the specifics of databases with a small number of records. The techniques selected by the author are proposed, which helped to break the deadlock in calculations, i.e., to get results much worse than expected. The need to apply methods to improve the accuracy of forecasts and the accuracy of classification was caused by a small amount of analysed data. This paper is not a review of popular methods of machine learning and data mining; nevertheless, the collected and presented material will help the reader to shorten the path to obtaining satisfactory results when using the described computational methods
... The improvement of the model quality by using a PCA has already been demonstrated in other studies (see, for example, [54]). By performing PCA, the 379 features were transformed into 46 principal components, which explain most of the variance of the original features. ...
Increasing time and cost pressure are forcing companies to plan and design their manufacturing systems with fuzzy product data in order to keep pace with shorter and agile product development cycles. This development is accompanied by a multitude of requirements for the planning of the respective manufacturing equipment. Many of these requirements are associated with a high degree of uncertainty and can thus only be specified vaguely in early planning phases, eventually leading to costly, originally unforeseen changes to the specified type of equipment. This consequently underlines the need for a predictive evaluation of uncertainty, or changeability in this context, to enable a more efficient planning approach based on quantified levels of uncertainty. In this paper, the authors therefore present the results of a systematic literature review on the evaluation of changeability in manufacturing systems design with special focus on uncertainty. Most importantly, standardized methods are not available yet; hence, product development as well as project management approaches are frequently adapted to manufacturing systems design. Furthermore, FMEA and fuzzy-logic-based methods are promising techniques for the assessment of uncertainty as a key element of changeability. Concluding, the paper discusses how the findings could support the development of a holistic approach to identify and predictively evaluate uncertainty in order to use it as a decision-making factor for the application of agile planning methods, thus contributing to better decision-making and higher achievement levels of project targets in industrial practice.
... On peut, par exemple, les normaliser ou standardiser : autres prétraitements tels que la PCA[WEG87] sont également envisageables. Ces prétraitements permettent d'éviter les effets indésirables des valeurs aberrantes sur la précision[JS11] ou d'extraire les informations importantes de données volumineuses ce qui peut contribuer à la précision du réseau, mais surtout accélérer les temps de calculs[MH+08]. ...
Aujourd’hui les données sont de plus en plus externalisées et réutilisées que ce soit dans la santé ou de façon plus générale. L’intérêt de ces techniques est de réduire les coûts associés aux infrastructures informatiques, mais également de permettre un accès simplifié aux données tout en laissant la possibilité de les traiter grâce à des algorithmes de Big data. En effet, des algorithmes de machine learning appliqués sur ces données permettent d’aider le professionnel dans sa prise de décision. Il est néanmoins primordial de rappeler que les données personnelles donc en particulier celle appartenant au secteur de la santé sont des données sensibles. En conséquence de quoi, des réglementations nationales ou internationales imposent leur sécurisation. Cela passe par le fait d’assurer leur confidentialité au travers d’algorithmes de chiffrement ainsi que leur intégrité et traçabilité grâce au tatouage numérique. D’autre part, il est nécessaire de comprendre que les volumes de données utilisées sont très important. Ceci impose que les données et en particulier celle issues de l’imagerie sont compressées afin de réduire les coûts de stockage. C’est pourquoi dans cette thèse, nous nous intéresserons aux trois problématiques associées aux données que sont leur sécurisation, traitement et compression. L’objectif sera alors d’associer ces trois propriétés afin de fournir une solution globale permettant de traiter de façon sécurisée des données compressées. Au cours de ces trois années de thèse, nous avons commencé par nous intéresser aux algorithmes de sécurisation de données qui permettent de les traiter. Ces algorithmes plus connus sous le nom de cryptosystèmes homomorphes existent sous plusieurs formes. Le chiffrement additivement homomorphe permet un nombre limité de traitement, mais est relativement peu coûteuse. Le fully homomorphe augmente les possibilités de traitements au détriment des performances.Notre première solution propose d’allier un cryptosystème additivement homomorphe avec un système de concaténation qui permet une compression dans le domaine chiffré. Elle permet d’effectuer un traitement matriciel sécurisé des données. Cette solution s’inscrit dans le projet Followknee dans lequel l’objectif est de traiter de façon sécurisée des données issues d’une prothèse connectée sur un smartphone. Par la suite, nous avons entrepris d’observer l’influence de la compression sur la précision des modèles de machine learning. L’objectif de cette partie étant de voir s’il est possible de ne décompresser que partiellement des données et ainsi gagner en termes de coûts calculatoires avant de les utiliser dans de tels modèles. Suite à quoi,nous avons essayé de décompresser partiellement et de façon sécurisée des données grâce à des cryptosystèmes fully homomorphes. Ceci représente la première étape d’une chaîne de traitement plus complexe permettant de compresser des données puis de les chiffrer avant de les externaliser vers un cloud.Une fois externalisées, ces données pourront être partiellement décompressées afin d’être utilisées de façon sécurisée au travers d’algorithmes de traitement. Pour finir, nous avons étudié la possibilité d’utiliser un cryptosystème additivement homomorphe en combinaison de calcul multipartite afin de sécuriser l’apprentissage d’un réseau de neurones artificiels.
... The improvement of the model quality by using a PCA has already been demonstrated in other studies (see, for example, [54]). By performing PCA, the 379 features were transformed into 46 principal components, which explain most of the variance of the original features. ...
Quality assurance (QA) is an important task in manufacturing to assess whether products meet their specifications. However, QA might be expensive, time-consuming, or incomplete. This paper presents a solution for predictive analytics in QA based on machine sensor values during production while employing specialized machine-learning models for classification in a controlled environment. Furthermore, we present lessons learned while implementing this model, which helps to reduce complexity in further industrial applications. The paper’s outcome proves that the developed model was able to predict product quality, as well as to identify the correlation between machine-status and faulty product occurrence.
