This research highlights the importance of integrating machine learning algorithms with Geographical Information Systems (GIS) applications in the field of renewable energy by finding a suitable site for wind farms due to their importance in preserving the environment to achieve efficiency and cost-effectiveness and reduce the environmental impact of fossil fuel energy sources. Using GIS various factors affecting wind energy localization were processed and analyzed including natural, socio-economic and environmental criteria. Ensemble learning of four supervised machine learning algorithms (Random Forest, K-Nearest Neighbor, Support Vector Machines, Naive Bayes) was used to classify suitable and unsuitable data representing geo-referenced points on the ground with three criteria for each site (wind speed, elevation and slope). The results of the algorithms varied in terms of accuracy and variance, then the results were collected, and the intersection between them was found so that the location classification would be agreed upon in the results of the algorithms used. The aim of using this technique is to reduce the error, increase the accuracy and avoid the bias or variance present in individual models. Accuracy of the algorithms result was respectively (K-Nearest Neighbor, Random Forest, Support Vector Machines, Naive Bayes) (93.022%, 93.018%, 95.095%, 89.553%). The final result is a map using GIS showing the suitable and unsuitable sites of wind farms in the study area (Türkiye) has been chosen as a study area in the research due to several factors that make it suitable for wind energy projects, including its geographical location, which gives it great climatic and terrain diversity, as it is surrounded by seas (Black Sea, Aegean Sea, and Mediterranean Sea), which leads to the activity of seasonal and continuous winds, which contributes to the activity of seasonal and permanent winds. Its drive to develop investment in renewable energy due to economic and population growth has increased the demand for energy and consequently the development of renewable and sustainable energy sources. This research contributes to supporting the global transition to sustainable energy by providing a new methodology for integrating multiple technologies to support a sustainable energy future.

Glaucoma, a major reason for incurable blindness globally, is still a significant public health issue. Existing diagnostic techniques are extremely clinician-reliant and time-consuming, resulting in undue delays in detection and treatment. This work proposes a new intelligent approach using Deep Convolutional Neural Networks (DCNNs) for the detection of glaucoma and optic disc segmentation from ophthalmic medical imaging. The suggested methodology includes preprocessing retinal fundus images to improve quality, followed by feature extraction through a DCNN structure optimized for glaucoma detection. Segmentation of the optic disc is performed through the VGG-19 model. Performance metrics confirm the efficiency of the suggested approach. The constructed DCNN model proves 98.69% accuracy in differentiating between glaucomatous and non-glaucomatous eyes, far exceeding conventional methods (85–90%). The model has a 95.18% recall, which means that the majority of actual glaucoma cases are identified correctly, and an F1-score of 96.84%, which reflects a good balance between precision and recall. In addition, with an AUC-ROC of 97.63%, the model is able to distinguish glaucomatous eyes well. Experimental results validate that the proposed approach improves accuracy and efficiency compared to current methods. This work contributes to the development of automated glaucoma detection algorithms with potential clinical uses in ophthalmology and medical imaging.

Abstract Network security in smart cities has become a key problem in the rapid development of computer networks over the past few years. Intrusion detection systems play a fundamental part in the integrity, confidentiality, and resource accessibility among the multiple network security policies. The classification of the genuineness of packets is main object of the presents research work, the soft computing has applied to classify the genuineness of packets. The complexity of soft computing is greatly reduced if the numbers of features in a dataset are reduced. Managing and analysis of the dimensionality reduction is novelty of the proposed model. The existence of uncertainty and the imprecise nature of the intrusions appear to create suitable fuzzy logic systems for such structures. The neural‐fuzzy algorithm is one of the effective methods that incorporate fuzzy logic systems into adaptive and analysis capacities. In this research work, soft computing fuzzy logic system is proposed to enhance network security through intrusion detection. Three network datasets are demonstrated to test and estimate the proposed system. Feature selection has used to remove irrelevant features from entire network data which are obstacle classification processes. The Information Gain method was applied to select importance features for detection intrusion. Adaptive Neuro‐Fuzzy Inference System (ANFIS) is further used to process the significant features of the classification network data as normal or attacks packets. Two functions named Jang's Neuro‐fuzzy and faster‐scaled conjugate gradient (SCG) based on the ANFIS system. Obviously, the experimental results demonstrate the proposed system has attained higher precision in detecting normal or attack. The experimental results have suggested that the proposed system results are better in accuracy and time process for classification compared with the existing models. The Overall Results show that the proposed system can be able to detect various intrusions efficiently and effectively.

The Internet of things (IoT) allows billions of computer and networking devices to be interconnected. Digital entities such as sensors, Radio-frequency identification (RFID), Internet and localization technologies enable everyday objects to be transformed into intelligent objects, which can communicate with one another. The built-in sensors in intelligent objects track and collect various types of data about equipment, the environment and human social life. Although the IoT is useful, safety susceptibility is a major concern. There are universal and continuous interconnections between people, devices, sensors and services. Even if the security system is well built, smartly configured, implemented effectively and managed correctly, it must be human-made and are not resistant to security threats. Therefore, the conception of cyber security solutions requires a human element. To fight attacks, this paper presents an Intrusion Detection System (IDS) based on Smart Techniques such as k-nearest neighbors (KNN) and Random Forest (RF) algorithms which take data as input for spotting harmful activities. The methodology of IDS was created for identifying malicious activities of network traffic utilizing a set of features. Designing efficient IDS involves a data source that incorporates a collection of attributes for evaluating its performance while identifying normal and attack instances. This is one of the challenges that we strive to address in this study. Based on KNN and RF algorithms has presented a collection of features for spotting harmful activities that take advantage of IoT applications to utilize these procedures. The data sources are taken from the UNSW-NB15 source files.

Credit card is one of the modern payment methods widely spread all over the world. It provides excellent facilities in purchasing as well as selling operations. However, it suffers from fraud problems, causing considerable economic losses to banks, institutions, and individuals, amounting to billions of dollars annually. That has made great interest in finding systems and means with outstanding capabilities to confront fraud, whose patterns in addition to methods are increasing dramatically. One of the most prominent techniques used by researchers in this field is Machine Learning (ML) techniques. In this paper, we proposed some of the classification ML algorithms such as Logistic regression(LR), Linear Discriminant Analysis (LDA), and Naïve Bayes(NB), additionally, the boosting algorithm XGBoost to create models capable of detecting fraud. The dataset from Kaggle. We used performance metrics such as accuracy, precision, f1, recall, AUC confusion matrix to evaluate the models' performance. The XGBoost model presented the best results compared to other models.

Mobile Ad hoc Networks (MANET) provide a service in one way or another through network applications in several diverse areas due to the multiple characteristics of MANET, such as high-level mobility, decentralization of nodes, and physical insecurity. The Intrusion Detection System (IDS) function detects threats and stops them before they disrupt the network. Supervised machine learning techniques are easy to implement and comprehend because they use the knowledge and experience gathered from the available data to classify network nodes into normal or malicious. To enhance security in MANET networks in this paper, we proposed a security detection system employing classify misbehavior of nodes depending on supervised machine learning techniques such as Random Forest (RF) and Naive Bayes(NB) techniques. Using a network simulator-2 (ns-2) simulation was made to produce a data set, explained Our experimental results for our proposed system have good accuracy for detecting intrusion in the system. The RF was more efficient.

Abstract Breast cancer can be effectively detected and diagnosed using the technology of digital mammography. However, although this technology has been rapidly developing recently, suspicious regions cannot be detected in some cases by radiologists, because of the noise or inappropriate mammogram contrast. This study presents a classification of segmented region of interests (ROIs) as either benign or malignant to serve as a second eye of the radiologists. Our study consists of three steps. In the first step, spherical wavelet transform (SWT) is applied to the original ROIs. In the second step, shape, boundary and grey level based features of wavelet (detail) and scaling (approximation) coefficients are extracted. Finally, in the third step, malignant/benign classification of the masses is implemented by giving the feature matrices to a support vector machine system. The proposed system achieves 91.4% and 90.1% classification accuracy using the dataset acquired from the hospital of Istanbul University in Turkey and the free Mammographic Image Analysis Society, respectively. Furthermore, discrete wavelet transform, which produces 83.3% classification accuracy, is applied to the coefficients to make a comparison with the SWT method.

Abstract: In this study, an automatic three‐dimensional computer‐aided detection system for colonic polyps was developed. Computer‐aided detection for computed tomography colonography aims at facilitating the detection of colonic polyps. First, the colon regions of whole computed tomography images were carefully segmented to reduce computational burden and prevent false positive detection. In this process, the colon regions were extracted by using a cellular neural network and then the regions of interest were determined. In order to improve the segmentation performance of the study, weights in the cellular neural network were calculated by three heuristic optimization techniques, namely genetic algorithm, differential evaluation and artificial immune system. Afterwards, a three‐dimensional polyp template model was constructed to detect polyps on the segmented regions of interest. At the end of the template matching process, the volumes geometrically similar to the template were emhanced.

In this paper, a novel type of equalization technique, called "double decision feedback equalizer" (DDFE) is applied to multilevel turbo codes (MLTC) to improve BER performance, and the entire system is called "multilevel turbo equalization" (MLTEQ). The parallel input data sequences are encoded at each level by turbo encoders, and then the coded data sequences are mapped to M-PSK signals, where M depends on the level quantity. After these modulated signals are passed through severe ISI and fading channels, the corrupted signals are equalized by an innovative iterative double decision feedback equalization technique, which uses an adaptive LMS algorithm to estimate the channel taps with a new equalization design. Then, the equalized signals are sent to the turbo decoders. The performance of new proposed MLTEQ system is investigated under non-frequency selective fading and frequency selective fading channels. As an application, two level turbo codes are simulated using 4-PSK modulation over AWGN, Rician, Rayleigh and Proakis B channels with 800 frame sizes. The simulation results show that satisfactory performance improvement is obtained with the proposed system over severe ISI and non-frequency selective fading channels.

In this paper, a blind maximum-likelihood channel estimation algorithm is developed for turbo trellis-coded/continuous-phase modulation (TTC/CPM) signals propagating through additive white Gaussian noise (AWGN) and Rician fading environments. We present CPM for TTC signals, since it provides low spectral occupancy and is suitable for power- and bandwidth-limited channels. Here, the Baum-Welch (BW) algorithm is modified to estimate the channel parameters. We investigate the performance of TTC/CPM for 16-CPFSK over AWGN and Rician channels for different frame sizes, in the case of ideal channel state information (CSI), no CSI, and BW estimated CSI.

In this paper, to improve bandwidth efficiency and error performance, partial response signaling (PRS) and trellis coded modulation (TCM) are combined together and named as QPR-TCM. The emphasis in this paper is on the jitter and error performance of QPR-TCM and neural network equivalent models in the fading environment with ideal channel state information (CSI) and no side information exist on the phase noise process. Analytical upper bounds are derived using Chernoff bounding technique, combined with the modified generating functional approach. One interesting result is that the bit error probability upper bounds of the proposed scheme is better than 4QAM-TCM for SNR values greater than a threshold, which have the best error performance till now. Neural network models of QPR-TCM schemes are introduced for the first time in the literature for fading channels and the bit error event curves are derived using the back propagation algorithm. It is shown that neuro-computing results confirm the analytical approach.

In order to improve the bandwidth efficiency and error performance, partial response signaling (PRS) and trellis coded modulation (TCM) are combined together. This method is named QPR-TCM. Whereas previous contributions have considered the performance of TCM transmitted over an additive white Gaussian noise (AWGN) channel, the emphasis in this paper is on the performance of trellis coded M-ary QAM combined with PRS in the fading environment. We only consider the case where interleaving/deinterleaving is employed to further combat the fading and assume that the ideal channel state information (CSI) is available. Analytical upper bounds are derived using the well known transfer function bounding technique, the bandwidth efficiencies are computed and the numerical results for several practical schemes are compared to each other. One interesting result is that when the combined system is used on a Rician fading channel, the bit error probability upper bounds of the proposed 6QPR-TCM and 9QPR-TCM systems are better than their counterparts the 4QAM-TCM systems with 2 and 4 states, respectively, for SNR values greater than a threshold, which have the best error performances in the literature.>

In order to improve the bandwidth efficiency and error performance, partial response signalling and trellis coded modulation are considered together for QAM. A new receiver for a data transmission system employing quadrature partial response trellis coded modulation (QPR-TCM) is investigated and upper bounded on mobile satellite fading channels. This scheme is also tight lower bounded on Rician fading channel regarding the length of the shortest error event path, the product of branch distances along that path.>

To improve both bandwidth efficiency and error performance, partial response signaling and trellis coded modulation are considered together for QAM. A new receiver for a data transmission system employing combined partial response/trellis coded modulation (QPR-TCM) is investigated. Combined 6QPR-TCM and 42QPR-TCM systems are introduced. Simulation studies in the ISI environment are done. In a colored noise environment, the error probability of 6QPR-TCM is analytically lower bounded and compared to the classical 4QAM-TCM. An optimal (QPR-TCM) scheme is proposed which improves both jitter and error response in a colored noise environment.>

https://doi.org/10.1109/pimrc.1991.571481