F IGURE 7: A ) TWO - CHANNEL FILTER BANK IMPLEMENTATION OF DWT APPLIED TO THE DATA VECTOR X . T HE BLOCKS G [.] AND H [.] REPRESENT THE HIGH PASS FILTER AND THE LOW PASS FILTER RESPECTIVELY . T HE IMPLEMENTATION CAN BE CARRIED OUT IN MORE RESOLUTION BY SUCCESSIVELY SPLITTING THE LOW PASS FILTER . B ) ECG SIGNAL APPROXIMATION AND DETAIL COEFFICIENTS AT LEVELS 1, 2, 3, 4&5. 22 F IGURE 8: A CLASSIFICATION PROBLEM . .......................................................................................... 26
The impact of innovated Neuro-FuzzySystem (NFS) has emerged as a dominant tech- nique for addressing various difficult research problems in business. ANFIS (Adaptive Neuro-Fuzzy Inference system) is an efficient combination of ANN and fuzzy logic for modeling highly non-linear, complex and dynamic systems. It has been proved that, with proper number of rules, an ANFIS system is able to approximate every plant. Even though it has been widely used, ANFIS has a major drawback of computational complex- ities. The number of rules and its tunable parameters increase exponentially when the numbers of inputs are large. Moreover, the standard learning process of ANFIS involves gradient based learning which has prone to fall in local minima. Many researchers have used meta-heuristic algorithms to tune parameters of ANFIS. This study will modify ANFIS architecture to reduce its complexity and improve the accuracy of classification problems. The experiments are carried out by trying different types and shapes of mem- bership functions and meta-heuristics Artificial Bee Colony (ABC) algorithm with ANFIS and the training error results are measured for each combination. The results showed that modified ANFIS combined with ABC method provides better training error results than common ANFIS model.
provide us with a great deal of information on the normal and pathological physiology of heart activity. Thus, ECG is an important non-invasive clinical tool for the diagnosis of heart diseases. For more than four decades, computers have been used in the classification of the ECG resulting in a huge variety of techniques [1,2] all designed to enhance the classification accuracy to levels comparable to that of a ‘gold standard’ of expert cardiology opinion. Included in these techniques are multivariate statistics, decision trees, fuzzy logic, expert systems
Over the years, many different methods and/or techniques have been developed for detecting and extracting the FECG signal. Blind Source Separation (BSS) is one of the methods that were recently investigated to measure the FECG signal. This is because the method involves the separation and the estimation of the original source of waveforms from a sensor array, without knowing the transmission channel characteristics and sources which may include diaphragm and uterus besides the FECG and the maternal electrocardiogram (MECG) signals (Lathauwer et al, 2000). This method fails in precise extraction of the FECG signal, since the contamination from the MECG signal is easily recognizable in the extracted FECG signal (Jang, 1993). Besides the BSS – base methods, various non-invasive signal processing techniques have been developed, such as fuzzy – logic, polynomial networks as well as wavelet theory based methods which can also be used to remove any ECG – noise signal from the FECG signal etc.
an accuracy of 91.83%, sensitivity of 91.83%, and specificity of 98.36%. The use of fuzzy member- ship function can cope with data that are not cla- ssifiable in the one against one SVM method to deliver performance values of accuracy, sensitivi- ity, and specificity higher than SVM method. This research is only using one lead of ECG signals. Some arrhythmia disorders common feature but is different leads, such as abnormalities in Left Bun- dle Branch Block Beat (LBBB), and Beat Right Bundle Branch Block (RBBB). Further research can be done using two leads to improve the accu- racy of the arrhythmia classificationsystem. References
NN can perform the necessary transformation and clustering operations automatically and simultaneously in parallel mode. An artificial NN (ANN) is able to abstract the distinctions between normal and premature ventricular contraction (PVC) patterns during the training sessions even though such distinctions may not be apparent. Second, a NN can recognize complex and nonlinear feature groups in the hyperspace. This is a distinct advantage over many conventional techniques. Third, a NN is massively parallel in nature and is capable easily operate in real-time mode . Several studies have presented the performance of NN systems for detection and recognition of abnormal ECGs. However, little work has been devoted to reducing the network processing time while maintaining required classification accuracy. The processing speed of basic ANNs can be increased by using their parallelization and realization on microcomputer systems in order to design real-time ECG monitoring systems  – . Efficient ANN for ECG monitoring system can be designed based on Adaptive Resonance Theory (ART). ART is a cognitive and neural theory of how the brain autonomously learns to categorize, recognize, and predict objects and events in a changing world. An important advantage of the ART is fast and stable learning . The algorithms of the ART perform as offline as online learning, they are capable to be very scalable for large scale datasets. ART has a possibility to correctly process noise data. Many traditional hierarchical and partition clustering algorithms fail to deal with large and high dimensional databases, especially in real-time situations. ART based methods are able to solve efficiently such problems.
recognition system must first capture digital image of Marathi handwritten numerical. Before attempting to classify the numerals, some preprocessing image might be necessary. For classification apply neural fuzzysystem. Two intensity values are available in binary image. These values are Black and White. We are use zero for Black and one for white. Thus the color of the character is White and the background is black. Preprocessing techniques are needed on color, grey-level or binary document images containing numerical and/or
There has been quite a number of software developed for the purposes of classification or clustering of bioinformatics data. As for this project, it is intended to apply a new classification technique in Adaptive NeuroFuzzy Inference System (ANFIS) model. After intensive literature reviews on various classification techniques, conventional or recurrent, fuzzy C means is selected to be used. And this technique will be integrated with ANFIS (Adaptive NeuroFuzzySystem) model by Jang, 1993 . Generally, there are 6 stages involved in the proposed model which are starting from the data input to output and it was developed using MATLAB software. This model requires data management before the classification and neurofuzzy model phase. Detection phase will follow once the output from the classification technique is done and lastly, performance analysis based on the result will be done at the end of the development phase. A layout of the development phase can be summarized as block diagram below.
Cardiac arrhythmias are one of the reasons for high mortality rate. The study of ECG pattern and heart rate variability in terms of computer-based analysis and classification of diseases can be very helpful in diagnostics . This thesis falls under the field of CI. Applications range from adaptive learning to speech recognition. The field draws from biological concepts such as the brain and the physiological decisions of organisms. Breakthroughs in applying CI to medical diagnosis have shown to be successful in the past and are thus important to heart signal classification . This type of data processing could be effectively applied to other biometrics such as electroencephalograms (EEGs) or electromyography (EMGs) to detect the existence of abnormalities in the brain or muscles respectively .
The separability between healthy and those with arrhythmia patients (Fig. 3a), also between healthy and those with AF (Fig. 3b), is due to the information obtained from the voltage variability and the morphology of the ECG signal, i.e., in addition to the frequency modula- tion information used in the heart rate variability (HRV), we added the amplitude modulation. Thus, to classify different types of arrhythmias, we use two modulation information of the ECG signal (frequency and amplitude). In addition, the value in the right part of the image (cross red) corresponds to a single beat that is located at the beginning of the ECG signal. Of course, the 1% rejection window used in this work was not large enough to exclude this beat.
Due to the presence of imprecise input information, ambi- guity or vagueness in input data, overlapping boundaries among classes, and indeﬁniteness in deﬁning features some uncertainties can still arise at any stage of a data classiﬁcation system. The fuzzy set theory [12–14] as a generalization of the classical set theory is very ﬂexible in handling different aspects of uncertainties or incompleteness about real life situations. In a fuzzysystem the features are associated with a degree of membership to different classes. Both NNs and fuzzy systems are very adaptable in estimating the input–output relation- ships. Neural networks deal with numeric and quantitative data while fuzzy systems can handle symbolic and qualitative data. Neuro-fuzzy hybridization leads to a crossbreed intelli- gent system widely known as Neuro-fuzzysystem (NFS) [15,16] that exploits the best qualities of these two approaches efﬁciently. The hybrid system unites the human alike logical reasoning of fuzzy systems with the learning and connected- ness structure of neural networks by means of fuzzy set theory based approach.
In this study we presented a neuro-fuzzysystem for clas- sification of asthma and chronic obstructive pulmonary disease (COPD). According to GINA and GOLD guide- lines we defined fuzzy rules and neural network para- meters. ANN of system was trained on more than one thousand medical reports obtained from database of the company CareFusion. Implemented neuro-fuzzysystem was validated on 455 patients by physicians from the Clinical Centre University of Sarajevo. All patients were separated into two groups, healthy and diseased. Diseased subjects were separated into two subgroups, asthmatics and COPD patients. Out of 170 asthmatic patients, neuro-fuzzysystem correctly classified 99.41% of patients. In addition, out of 248 COPD patients 99.19% were correctly classified. The system was 100% correct on 37 patients with normal lung function. Based on our neuro-fuzzysystem we obtained sensitivity of 99.28% and specificity of 100% in asthma and COPD classification. These results have been achieved due to the fact that in our neuro-fuzzysystem are also implemented all recom- mendations of GINA and GOLD necessary for classifica- tion of asthma and COPD. Also, as shown in the results, in the process of establishing the final diagnosis, com- plete dynamic assessment of the patient is obtained, as opposed to the solution that provides a static assessment of the patient.
ANFIS is a kind of hybrid of neural network and fuzzy logic and is based on Takagi–Sugeno fuzzy inference system. In ANFIS, we combine both the learning capabilities of a neural network and reasoning capabilities of fuzzy logic in order to give enhanced prediction capabilities . Since it integrates both neural networks and fuzzy logic principles, it has potential to capture the benefits of both in a single framework. Its inference system corresponds to a set of fuzzy  IF–THEN rules that have learning capability to approximate nonlinear functions. Hence, ANFIS is considered to be universal approximator. The ANFIS model is very suitable and can generate excellent classiﬁcation results provided that the right type and number of Membership Functions (MFs) are used in the classiﬁcation task .In the classification  two different classification techniques are employed: an artificial neural network-based classifier and a hybrid ANFIS classifier. A neural classifier can learn from data, but the output does not lead itself naturally to interpretation. An ANFIS classifier is based on a three-layer feed-forward neural network and combines the merits of both neural and fuzzy classifiers while overcoming their drawbacks and limitations. The developed Adaptive NeuroFuzzy Inference System (ANFIS) classifier exhibits high levels of accuracy, consistency and reliability, with acceptably low computational time and is a promising new development in the field of diagnosis of cancer.
After acquiring the output from the develop diagnostic system for renal or kidney cancer by using adaptive neuro- fuzzy inference system(ANFIS), now the specialist doctors or expert detects the obtained output by comparing it with the actual or target output. Here, the target output is that output which is evaluated by the expertise for a particular inputs and same inputs are given to diagnostic system. After the comparing between obtained and target results, it is noticed that the developed diagnostic system for renal cancer gives the similar output as the target outputs. The performance of the diagnostic system for the renal cancer is calculated by considering the various parameters. These parameters are sensitivity, specificity, precision and classification accuracy. The table 1 shows the performance of developed diagnostic system for renal cancer by using adaptive neuro-fuzzy inference system by considering the various parameters.
Neural networks and fuzzy logic are two totally inspired concepts of human reasoning. Although these two numerical models are different in terms of structure, they have many things in common. Indeed, the use of these techniques does not require mathematical model defined as the resolution is based on the numerical values of inputs and outputs for neural networks and logic of the system in the case of fuzzy logic. In addition, the results are generally uncertain. However these techniques can perform additional tasks. Fuzzy logic provides knowledge with a certain degree of uncertainty (or accuracy). On the other hand, neural networks can model and replicate human learning, where the idea of combining these two techniques to create an artificial intelligent system that provides us a resolution of problems as close as that of the human being.
Big data and cloud computing technology appeared on the scene as new trends due to the rapid growth of social media usage over the last decade. Big data represent the immense volume of complex data that show more details about behaviours, activities, and events that occur around the world. As a result, big data analytics needs to access diverse types of resources within a decreased response time to produce accurate and stable business experimentation that could help make brilliant decisions for organizations in real-time. These developments have spurred a revolutionary transformation in research, inventions, and business marketing. User behaviour analysis for classification and prediction is one of the hottest topics in data science. This type of analysis is performed for several purposes, such as finding users ’ interests about a product (for marketing, e-commerce, etc.) or toward an event (elections, championships, etc.) and observing suspicious activities (security and privacy) based on their traits over the Internet. In this paper, a neuro-fuzzy approach for the classification and prediction of user behaviour is proposed. A dataset, composed of users ’ temporal logs containing three types of information, namely, local machine, network and web usage logs, is targeted. To complement the analysis, each user ’ s 360-degree feedback is also utilized. Various rules have been implemented to address the company ’ s policy for determining the precise behaviour of a user, which could be helpful in managerial decisions. For prediction, a Gaussian Radial Basis Function Neural Network (GRBF-NN) is trained based on the example set generated by a Fuzzy Rule BasedSystem (FRBS) and the 360-degree feedback of the user. The results are obtained and compared with other state-of-the-art schemes in the literature, and the scheme is found to be promising in terms of classification as well as prediction accuracy.
Entropy values are the features and are given to the classifiers where we get the result as identified or rejected. The performance was analyzed using correct classification percentage (CCP) which is defined as the number of correctly identified images by total number of images tested. The CCP for the non trained images was found to be 70% using ANFIS and in the case of SVM, it was found to be less, as shown in Table 2. Results can be further improved by using adaptive lifting scheme for feature extraction technique.
their potential in assessing meat quality. In recent years, spectral imaging (i.e., hyperspectral and multispectral) has been also considered as an alternative tool for safety and quality inspection of various agricultural products . This technique integrates the conventional imaging and spectroscopy technique to attain simultaneously both spatial and spectral information from the target product. The “mechanism” of these approaches is based on the assumption that the metabolic activity of micro-organisms on meat results in biochemical changes, with the simultaneous formation of metabolic by-products, which may contribute to the spoilage phenomenon. The quantification of these metabolic activities corresponds to a unique “signature”, providing thus information about the type and rate of spoilage . The huge amount of information provided by analytical sensors/devices requires an advanced data analysis approach. This has been achieved through the integration of modern analytical platforms with computational and chemometric techniques . Multivariate statistical analyses (e.g., partial least square (PLS) regression, discriminant function analysis (DFA), cluster analysis) have resulted in the development of decision support systems for timely determination of safety/quality of meat products . Considering that microbial meat spoilage is a complex process, which involves growth of microorganisms during storage, their spectra contain highly non-linear characteristics. Hence, linear-based techniques might not provide a complete solution to such complex identification/classification problem . Neural networks (NNs) have gained much interest in predictive engineering and quantitative modelling due to their flexibility and high accuracy as compared to other modelling techniques (e.g., statistical models). In comparison to other NN-based application areas, the field of food science is still in an early development stage. Recently, advanced NN algorithms have shown promising results in applications such as growth parameter estimation of microorganisms . NNs usually require a large number of neurons for solving the majority of approximation problems and are prone to dimensionality problems, as each single neuron- node cannot define a multi-dimensional hyper-sphere of the input domain. Although fuzzy logic systems, provide such input space mapping, they do not have learning ability, thus it is difficult to analyse complex systems without prior and accurate knowledge on the system being analysed .
In our research, leaf images are the source of information and are being used to train the neuro-fuzzy classifier. The common shape features of the leaf are the length, width, area and perimeter. These features can be extracted by processing the Canny edge detection, binarization and others. Some images are divided into sets of ratio segments to improve the accuracy of the system. The shape features were also used by (Chaki & Bhattacharya, 2015). Chaki & Bhattacharya (2015) proposed the neuro-fuzzy classifier, and in order to test the accuracy of the system, they compared it with the Neural Network and k-Nearest Neighbour classifiers. The study analysed 640 images of leaves to classify them into 32 predefined classes. The leaves were subjected to the image pre-processing steps. Firstly, the scale and the orientation of the images transformed in same standardization and later the computation of the shape features is done. Images were converted into binary form. The angle of the major axis, which is the length of the leaf was oriented/placed horizontally. Next, the background was cropped until the leaf fits within the boundary of the rectangle. The images also were scaled into pre-defined sizes called “segments” according to the aspect ratio values of the leaves. Based on these workflows, we adapted the similar steps and further adaption is discussed in the next section.
component analysis (PCA) technique. These features are used to train the neurofuzzy classifier. The neurofuzzy classifier is used for classification is the Adaptive Network basedFuzzy Inference system(ANFIS).The developed neurofuzzy classifier is tested for classification of different brain MRI samples. Thus, the proposed work emphasizes on development of Neural Network and Fuzzy logic based method for the classification of MRI brain images. The block schematic diagram shown in figure 1 is the proposed architecture for classification of MRI brain images.