Papers by Dr. Fuad Alkoot
International Journal of Data Mining and Bioinformatics, 2016
Automated autism detection is needed to facilitate urgently required therapy. However, contrary t... more Automated autism detection is needed to facilitate urgently required therapy. However, contrary to cancer, autism detection using genetic data has not attracted much attention. In this paper, we investigate autism detection using machine learning techniques. The main goal is to test whether genetic data with machine learning tools can result in an abbreviated and accurate instrument for classification of autism. For this, a system comprising four stages is proposed, where at each stage, we experiment with different feature reduction, classification and combination methods to find if it is possible to detect autism. The experimental results show that our classifier-based system can achieve optimum accuracy of early screening. We achieved optimum accuracy when examined on independent and unseen test data. The optimum performance was mostly achieved using a three-layer back-propagation neural network classifier combined using the feature selection-based combiner. This was achievable only when the data dimensionality was reduced using our proposed feature selection method. The maximum number of features varied for the different chromosomes and ranged between 150 and 500.
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2003
Amidst the conflicting experimental evidence of superiority of one over the other, we investigate... more Amidst the conflicting experimental evidence of superiority of one over the other, we investigate the Sum and majority Vote combining rules in a two class case, under the assumption of experts being of equal strength and estimation errors conditionally independent and identically distributed. We show, analytically, that, for Gaussian estimation error distributions, Sum always outperforms Vote. For heavy tail distributions, we demonstrate by simulation that Vote may outperform Sum. Results on synthetic data confirm the theoretical predictions. Experiments on real data support the general findings, but also show the effect of the usual assumptions of conditional independence, identical error distributions, and common target outputs of the experts not being fully satisfied.
Experimental analysis of machine learning methods to detect Covid-19 from x-rays
Journal of Engineering Research
Microarray Gene Expression Data Dimensionality Reduction Using PCA
World Academy of Science, Engineering and Technology, International Journal of Computer and Information Engineering, 2015
Improving the performance of the product fusion strategy
Proceedings 15th International Conference on Pattern Recognition. ICPR-2000
ABSTRACT Among existing classifier combination rules the most widely used are sum, product and vo... more ABSTRACT Among existing classifier combination rules the most widely used are sum, product and vote. Although product is more directly related to the compound class posterior probability, it does not perform well. Sum, which is derived under restricting assumptions, outperforms product, especially if the class aposteriori probability estimates are subject to high levels of noise. We establish the cause of product's degraded performance and propose a method to improve it. Tests on real and synthetic data demonstrate that the modified product has a number of advantages in relation to other rules that we experiment with
Classifier Combination as a Tomographic Process
Lecture Notes in Computer Science, 2001
A mathematical analogy between the process of multiple expert fusion and the tomographic reconstr... more A mathematical analogy between the process of multiple expert fusion and the tomographic reconstruction of Radon integral data is outlined for the specific instance of the combination of classifiers containing discrete data sets. Within this metaphor all conventional methods of classifier combination come, to a greater or lesser degree, to resemble the unfiltered back-projection of the constituent classifiers’ probability density
Using Classifiers to Predict Student Outcome at Higher Institute of Telecommunication
World Academy of Science, Engineering and Technology, International Journal of Computer and Information Engineering, 2016
Video-based biometric systems are becoming feasible thanks to advancement in both algorithms and ... more Video-based biometric systems are becoming feasible thanks to advancement in both algorithms and compu-tation platforms. Such systems have many advantages: improved robustness to spoof attack, performance gain thanks to variance reduction, and increased data qual-ity/resolution, among others. We investigate a discrimi-native video-based score-level fusion mechanism, which enables an existing biometric system to further harness the riches of temporarily sampled biometric data using a set of distribution descriptors. Our approach shows that higher order moments of the video scores contain discrim-inative information. To our best knowledge, this is the first time this higher order moment is reported to be effective in the score-level fusion literature. Experimental results based on face and speech unimodal systems, as well as multimodal fusion, show that our proposal can improve the performance over that of the standard fixed rule fu-sion strategies by as much as 50%. 1
We experiment with different fusion methods when bagging k-NN classifiers under various condition... more We experiment with different fusion methods when bagging k-NN classifiers under various conditions. Experiments with four types of bagging are made at four training set sizes, using two metrics. The aim is to find the conditions for an optimum bagging performance. Additionally we aim to find the best rule under the specified conditions. We compare the performance of the different fusion strategies under each condition. Fusion methods used are Sum, Modified Product (MProduct) [2], Vote and Moderation [1]. Results show that the performance depends on the data used, number of nearest neighbors (k), number of fused classifiers and size of training set. Over all the three rules derived from Product show a close performance, while Vote shows an opposite performance. Among the three rules Moderation either follows Sum or MProduct. Results indicate MProduct outperforms Sum at many instances. At some of these instances Sum did not outperform the single classifier while MProduct did. Moderati...
We experiment with bagging kNN classifiers using an optimal distance metric. The aim is to establ... more We experiment with bagging kNN classifiers using an optimal distance metric. The aim is to establish whether bagging kNN is useful when a better metric is used. We experiment on real world data sets, at different training set sizes. Our experiments also involve Modified bagging, which was proposed by us, to see the effect of prior knowledge on the bagging performance under the new distance measure. Results indicate the optimal metric improves the performance of bagging as well as the single classifier. Key-Words: bagging, fusion, classifier combining, nearest neighbor
Monitoring Wi-Fi radiation levels at residences in Kuwait: A field survey
2014 11th International Conference on Wireless Information Networks and Systems (WINSYS), 2014
Many are questioning the health effects of EMF radiation transmitted by mobile phone network or s... more Many are questioning the health effects of EMF radiation transmitted by mobile phone network or simply the Wi-Fi networks at public and private locations. High frequency radiation has been linked to various types of illnesses and recently has been categorised as class 2B carcinogen. The Wi-Fi source has been considered as safe due to the low emitted radiation levels compared to the standard limits. However, many are questioning the adequacy of these limits. Additionally, the technology has become very common in the majority of residences resulting in an accumulation of several networks in each house. This increases the amount of radiation and supports the possibly of increased hazard. In this study, we aim at documenting the state of radiation due to WLAN sources at 2.4GHz at various residences in Kuwait. Results indicate that apartments in building complexes suffer from higher radiation sources due to the higher population density at these buildings. Although houses suffer from a s...
We aim to find the effect of diversity on combiner performance. Three diversity measures are used... more We aim to find the effect of diversity on combiner performance. Three diversity measures are used to calculate the diversity of combined classifiers. We aim to identify which measure is closely related to the combiner performance. Three combiner types are used; Bagging and a conventional three classifier system, in which three classifier types are used; backpropagation neural network, bayesian and k-nearest neighbor classifiers. Additionally we experiment with a feature based combiner system proposed by Alkoot [13,14]. Results obtained on real data indicate the diversity measure to be higher for systems with higher classification rate, if it outperforms other classifiers by a large margin. Otherwise, if the performances of the compared systems are close the diversity measure may not be higher for the best system. On many occasions the diversity measures were not good indicators of system performance. On some instances we found that the more diverse system did not yield a better perf...
We investigate a number of parameters commonly affecting the design of a multiple classifier syst... more We investigate a number of parameters commonly affecting the design of a multiple classifier system in order to find when fusing is most beneficial. We extend our previous investigation to the case where unequal classifiers are combined. Results indicate that Sum is not affected by this parameter, however, Vote degrades when a weaker classifier is introduced in the combining system. This is more obvious when estimation error with uniform distribution exists.
DNA data have been used in forensics for decades. However, current research looks at using the DN... more DNA data have been used in forensics for decades. However, current research looks at using the DNA as a biometric identity verification modality. The goal is to improve the speed of identification. We aim at using gene data that was initially used for autism detection to find if and how accurate is this data for identification applications. Mainly our goal is to find if our data preprocessing technique yields data useful as a biometric identification tool. We experiment with using the nearest neighbor classifier to identify subjects. Results show that optimal classification rate is achieved when the test set is corrupted by normally distributed noise with zero mean and standard deviation of 1. The classification rate is close to optimal at higher noise standard deviation reaching 3. This shows that the data can be used for identity verification with high accuracy using a simple classifier such as the k-nearest neighbor (k-NN). Keywords—Biometrics, identity verification, genetic data...
International Journal of Information and Education Technology
Higher education institutes face many challenges that require robust and scientific solutions. Si... more Higher education institutes face many challenges that require robust and scientific solutions. Six-sigma process improvement methodology is an example of a scientific method that we aim to use to improve the outcome of an educational institute. Six sigma is implemented successfully in the services and manufacturing sectors but rarely in education. Our aim is to show that six-sigma can be used in an educational institute of higher education and to show how it can be implemented. We use this scientific method for process control in order to improve the student outcome. We implement the first two phases of the six sigma method. In the first phase we state the problem and in the second phase we collect data from real cases at the Higher Institute of Telecommunication and Navigation. The obtained results will lead to an improvement of skills and knowledge attained by graduates of the institute. Preliminary results show that it diagnosis if problems exist and sets the path for finding causes that we must deal with to obtain improvements.
Investigating machine learning techniques for the detection of autism
International Journal of Data Mining and Bioinformatics, 2016
Automated autism detection is needed to facilitate urgently required therapy. However, contrary t... more Automated autism detection is needed to facilitate urgently required therapy. However, contrary to cancer, autism detection using genetic data has not attracted much attention. In this paper, we investigate autism detection using machine learning techniques. The main goal is to test whether genetic data with machine learning tools can result in an abbreviated and accurate instrument for classification of autism. For this, a system comprising four stages is proposed, where at each stage, we experiment with different feature reduction, classification and combination methods to find if it is possible to detect autism. The experimental results show that our classifier-based system can achieve optimum accuracy of early screening. We achieved optimum accuracy when examined on independent and unseen test data. The optimum performance was mostly achieved using a three-layer back-propagation neural network classifier combined using the feature selection-based combiner. This was achievable only when the data dimensionality was reduced using our proposed feature selection method. The maximum number of features varied for the different chromosomes and ranged between 150 and 500.
Proceedings of the Third International Conference on Information Fusion, 2000
We investigate two distinct design strategies which we refer to as parallel and serial. In both c... more We investigate two distinct design strategies which we refer to as parallel and serial. In both cases we show that the proposed integrated design approach leads to improved performance.
Improving Product by Moderating k-NN Classifiers
Lecture Notes in Computer Science, 2001
ABSTRACT The veto effect caused by contradicting experts outputting zero probability estimates le... more ABSTRACT The veto effect caused by contradicting experts outputting zero probability estimates leads to fusion strategies performing sub optimally. This can be resolved using Moderation. The Moderation formula is derived for the k-NN classifier using a bayesian prior. The merits of moderation are examined on real data sets.
Energy and Power Engineering, 2015
According to surveyed literature, there may be a health hazard associated with extremely low freq... more According to surveyed literature, there may be a health hazard associated with extremely low frequency magnetic fields. This study aims at presenting a recent survey of this literature. It also aims at measuring magnetic field levels close to power transmission lines at inhabited areas in Kuwait to see if current levels are safe and to establish a database of 50 Hz magnetic field levels at inhabited areas. Measurements were made, according to the international standard procedures in winter, spring, summer and fall, and three times of a day. Four inhabited areas were surveyed. Results provide us with an independent view of the levels in the vicinity of power lines and houses. Results show that the highest level is measured in the summer, reaching 115 mG while the minimum level is measured in the fall. We found that some houses were at less than 50 meters distance from the edge of the transmission lines. Some houses were, as close as 22 meters to the line. We found that levels at the entrance of houses and at outdoor parking areas were mostly higher than 4 mG.
Series in Machine Perception and Artificial Intelligence, 2002
Uploads
Papers by Dr. Fuad Alkoot