Search

Accurate predictions of stock markets are important for investors and other stakeholders of the equity markets to formulate profitable investment strategies. The improved accuracy of a prediction model even with a slight margin can translate into considerable monetary returns. However, the stock markets' prediction is regarded as an intricate research problem for the noise, complexity and volatility of the stocks' data. In recent years, the deep learning models have been successful in providing robust forecasts for sequential data. We propose a novel deep learning-based hybrid classification model by combining peephole LSTM with temporal attention layer (TAL) to accurately predict the direction of stock markets. The daily data of four world indices including those of U.S., U.K., China and India, from 2005 to 2022, are examined. We present a comprehensive evaluation with preliminary data analysis, feature extraction and hyperparameters' optimization for the problem of stock market prediction. TAL is introduced post peephole LSTM to select the relevant information with respect to time and enhance the performance of the proposed model. The prediction performance of the proposed model is compared with that of the benchmark models CNN, LSTM, SVM and RF using evaluation metrics of accuracy, precision, recall, F1-score, AUC-ROC, PR-AUC and MCC. The experimental results show the superior performance of our proposed model achieving better scores than the benchmark models for most evaluation metrics and for all datasets. The accuracy of the proposed model is 96% and 88% for U.K. and Chinese stock markets respectively and it is 85% for both U.S. and Indian markets. Hence, the stock markets of U.K. and China are found to be more predictable than those of U.S. and India. Significant findings of our work include that the attention layer enables peephole LSTM to better identify the long-term dependencies and temporal patterns in the stock markets' data. Profitable and timely trading strategies can be formulated based on our proposed prediction model.

Heart disease is a serious worldwide health issue with wide-reaching effects. Since heart disease is one of the leading causes of mortality worldwide, early detection is crucial. Emerging technologies like Machine Learning (ML) are currently being actively used by the biomedical, healthcare, and health prediction industries. PaRSEL, a new stacking model is proposed in this research, that combines four classifiers, Passive Aggressive Classifier (PAC), Ridge Classifier (RC), Stochastic Gradient Descent Classifier (SGDC), and eXtreme Gradient Boosting (XGBoost), at the base layer, and LogitBoost is deployed for the final predictions at the meta layer. The imbalanced and irrelevant features in the data increase the complexity of the classification models. The dimensionality reduction and data balancing approaches are considered very important for lowering costs and increasing the accuracy of the model. In PaRSEL, three dimensionality reduction techniques, Recursive Feature Elimination (RFE), Linear Discriminant Analysis (LDA), and Factor Analysis (FA), are used to reduce the dimensionality and select the most relevant features for the diagnosis of heart disease. Furthermore, eight balancing techniques, Proximity Weighted Random Affine Shadowsampling (ProWRAS), Localized Randomized Affine Shadowsampling (LoRAS), Random Over Sampling (ROS), Adaptive Synthetic (ADASYN), Synthetic Minority Oversampling Technique (SMOTE), Borderline SMOTE (B-SMOTE), Majority Weighted Minority Oversampling Technique (MWMOTE) and Random Walk Oversampling (RWOS), are used to deal with the imbalanced nature of the dataset. The performance of PaRSEL is compared with the other standalone classifiers using different performance measures like accuracy, F1-score, precision, recall and AUC-ROC score. Our proposed model achieves 97% accuracy, 80% F1-score, precision is greater than 90%, 67% recall, and 98% AUC-ROC score. This shows that PaRSEL outperforms other standalone classifiers in terms of heart disease prediction. Additionally, we deploy SHapley Additive exPlanations (SHAP) on our proposed model. It helps to understand the internal working of the model. It illustrates how much influence a classifier has on the final prediction outcome.

Effective resource allocation has always been one of the serious challenges in wireless communication. A considerable number of machine type communication (MTC) devices in 5G with variable quality of service (QoS) aggravates this challenge even further. Existing Resource allocation schemes in MTC are usually considering signal to noise ratio (SNR), which provides preference to MTC devices based on distance rather than their QoS requirements. This paper proposes a resource allocation scheme with dynamic priorities for MTC devices with multiple radio access technologies (RATs). The proposed resource allocation scheme has two main parts namely medium access and resource allocation. The medium access leverages the broadcast nature of wireless signal and MTC devices' wait time to assign priorities using capillary band in a secure and integral way. At resource allocation, SNR, total induced transmission delay, and transmission-awaiting MTC devices are used to assign resources in the cellular band. The rumination of two-staged dynamic priorities in the proposed scheduling scheme brings significant performance improvements in outage and success probabilities. Compared to SNR-based schemes, the proposed mechanism performs well by expressively improving the outage and success probability by 20% and 30%, respectively.

Named data networking (NDN) is one of the future networking architectures that communicates content using names, instead of the node addresses. It uses a very simple pull-based communication mechanism to retrieve content by sending an Interest message, and the node that has the required content or producer node replies with the data message. In wireless networks, the interest is flooded in the network to find the data provider node. The directional diffusion method is used to pull further content from the provider node. Due to broadcast nature and without node addresses, interest flooding causes network congestion and wastes network resources, especially bandwidth and battery power. These resources have prime importance in the case of wireless sensor networks (WSNs) because all WSN nodes operate on battery and have limited bandwidth. In this paper, we propose an interest broadcast suppression scheme that considers interest holding time using the distance between forwarder and receiver of the interest, energy, angle, and distance from the beeline between consumer and the spatial region, to avoid broadcasting of unnecessary copies of Interest. The simulation results show that the proposed scheme mitigates the interest broadcast and conserves battery power of the wireless nodes compared with the state-of-the-art scheme in the domain.

IEEE 802.15.4 standard is specifically designed for a low-rate and low-processing Internet of things (IoT) applications and offers guaranteed time slots. A beacon-enabled IEEE 802.15.4 consists of a superframe structure that comprises of the contention access period and contention-free period. During contention-free period, nodes transfer their data using guaranteed time slots without any collision. The coordinator node receives data transmission requests in one cycle and allocates guaranteed time slots to the nodes in the next cycle. This allocation process may cause large delay that may not be acceptable for few applications. In this work, a novel superframe structure is proposed that significantly reduces guaranteed time slots allocation delay for the nodes with data requests. The proposed superframe structure comprises of two contention access periods and one contention-free period, where contention-free period precedes both contention access periods with reduced slot size. In addition, the knapsack algorithm is modified for better guaranteed time slots allocation by allowing more guaranteed time slots requesting nodes to send their data as compared to the IEEE 802.15.4 standard. The simulation and analytical results show that the proposed superframe structure reduces the network delay by up to 80%, increases contention-free period utilization up to 50%, and allocates guaranteed time slots up to 16 nodes in a single superframe duration.

Modern vehicles equipped with a large number of electronic components, sensors, actuators, and extensive connectivity, are the classical example of cyber-physical systems (CPS). Communication as an integral part of the CPS has enabled and offered many value-added services for vehicular networks. The communication mechanism helps to share contents with all vehicular network nodes and the surrounding environment, e.g., vehicles, traffic lights, and smart road signs, to efficiently take informed and smart decisions. Thus, it opens the doors to many security threats and vulnerabilities. Traditional TCP/IP-based communication paradigm focuses on securing the communication channel instead of the contents that travel through the network. Nevertheless, for content-centered application, content security is more important than communication channel security. To this end, named data networking (NDN) is one of the future Internet architectures that puts the contents at the center of communication and offers embedded content security. In this paper, we first identify the cyberattacks and security challenges faced by the vehicular CPS (VCPS). Next, we propose the NDN-based cyber-resilient, the layered and modular architecture for VCPS. The architecture includes the NDN's forwarding daemon, threat aversion, detection, and resilience components. A detailed discussion about the functionality of each component is also presented. Furthermore, we discuss the future challenges faced by the integration of NDN with VCPS to realize NDN-based VCPS.

Wireless communications and networking have been continually evolving to improve and be a part of our lifestyle. This uninterrupted development is due to many research projects and practices that are being carried out to improve the quality of services and applications supported by networking technologies. Initially, the plethora of computer networks research was designed to allow users to share the thoughts and facts using textual data through addressing devices. Meanwhile, we have witnessed that users play the role of both producers and consumers at the same time. These new emerging requirements gave birth to Cloud Computing (CC), Data Centric Networking (DCN), and other advancements in IEEE standards. Moreover, researchers also intended to redesign the networking architectures and protocols with the focus on content rather than the host. The resulting new architectures are Information Centric Networks (ICN) with various extensions like Content Centric Networks (CCN), Named Data Networks (NDN), Data-Oriented Network Architecture, etc.

Smart cities have been envisioned during the last decade. Moreover, various projects have been initiated to bring this concept into reality. Smart city is basically an emergence of the existing and new Information and Communications Technologies (ICT) to make our living standard more safe and digitized. Wireless communications, such as sensors, actuators, intelligent transportation systems, and smart grids, have played a vital role in the dissemination of information under the given circumstances. Similarly, it is hard to declare any city as a smart city without taking benefits from wireless sensor networks (WSNs). However, with new requirements and delay sensitive applications, the existing WSN requires significant alterations at several layers. In this paper, a new time division multiple access (TDMA)-based medium access control (MAC) protocol, called bitmap-assisted efficient and scalable TDMA-based MAC (BEST-MAC), is proposed for adaptive traffic in hierarchical WSNs that can be deployed in the smart cities. BEST-MAC is specifically designed to improve the quality control of such smart cities applications, where diverse traffic is required and loss or delay in data traffic is unacceptable. The main contributions of BEST-MAC include: 1) it uses small size time slots; 2) the number of those time slots is more than the number of member nodes; 3) knapsack algorithm is used to schedule time slots; and 4) short node address (1 B) is proposed to identify the member nodes. First two contributions of BEST-MAC handle adaptive traffic loads of all members in an efficient manner. The knapsack algorithm not only reduces the job completion time of a node but also minimizes the average packet delay with better link utilization. The short node address reduces the control overhead that improves the energy efficiency. The simulation results verify that the proposed BEST-MAC transmits more data with less delay and energy consumption compared with the existing MAC protocols.

Flying Ad hoc Network (FANET) is a new resource-constrained breed and instantiation of Mobile Ad hoc Network (MANET) employing Unmanned Aerial Vehicles (UAVs) as communicating nodes. These latter follow a predefined path called ’mission’ to provide a wide range of applications/services. Without loss of generality, the services and applications offered by the FANET are based on data/content delivery in various forms such as, but not limited to, pictures, video, status, warnings, and so on. Therefore, a content-centric communication mechanism such as Information Centric Networking (ICN) is essential for FANET. ICN addresses the problems of classical TCP/IP-based Internet. To this end, Content-centric networking (CCN), and Named Data Networking (NDN) are two of the most famous and widely-adapted implementations of ICN due to their intrinsic security mechanism and Interest/Data-based communication. To ensure data security, a signature on the contents is appended to each response/data packet in transit. However, trusted communication is of paramount importance and currently lacks in NDN-driven communication. To fill the gaps, in this paper, we propose a novel trust-aware Monitor-based communication architecture for Flying Named Data Networking (FNDN). We first select the monitors based on their trust and stability, which then become responsible for the interest packets dissemination to avoid broadcast storm problem. Once the interest reaches data producer, the data comes back to the requester through the shortest and most trusted path (which is also the same path through which the interest packet arrived at the producer). Simultaneously, the intermediate UAVs choose whether to check the data authenticity or not, following their subjective belief on its producer’s behavior and thus-forth reducing the computation complexity and delay. Simulation results show that our proposal can sustain the vanilla NDN security levels exceeding the 80% dishonesty detection ratio while reducing the generated end-to-end delay to less than 1 s in the worst case and reducing the average consumed energy by more than two times.

We propose two routing protocols for Terrestrial Wireless Sensor Networks (TWSNs): Hybrid Energy Efficient Reactive (HEER) and Multihop Hybrid Energy Efficient Reactive (MHEER) routing protocol. The main purpose of designing these protocols is to improve the network lifetime and particularly the stability period of the underlying network. In MHEER, the node with the maximum energy in a region becomes cluster head (CH) of that region for that particular round (or cycle) of time and the number of the CHs in each round remains the same. Our techniques outperform the well-known existing routing protocols: LEACH, TEEN, and DEEC in terms of stability period and network lifetime. We also calculate the confidence interval of all our results which helps us to visualize the possible deviation of our graphs from the mean value. We also implement sink mobility on HEER and MHEER. We refer to them as HEER-SM and MHEER-SM. Simulation results show that HEER-SM and MHEER-SM yield better network lifetime and stability region as compared to the counterpart techniques. We have also carried out simulations with 500 and 1000 nodes in the same field dimensions besides 100 nodes. Simulations prove that the proposed schemes show the same behavior with 500 and 1000 nodes; that is, HEER and MHEER are scalable as well.

Beacon or safety messages are broadcasted in vehicular ad hoc networks (VANETs) to disseminate network state or emergency incident information to other vehicles in the network. The freshness of information depends upon the frequent transmission of beacons. Similarly, to increase the awareness area or communicating with the distant nodes, beacons are disseminated with a high transmission power. However, increasing the beacon transmission power or rate has a negative effect on information communication efficiency because of the finite bandwidth of the wireless link. Therefore, different schemes have been proposed to individually control beacon's transmission power, transmission rate, or contention window at the MAC layer, or any combination of those, to achieve quality beacon communication in VANETs. The latter case is called hybrid adaptive beaconing schemes. In literature, there are many hybrid adaptive beaconing schemes that control multiple communication parameters to efficiently broadcast beacon messages in VANETs. In this paper, we explicitly survey and summarized various aspects of those schemes. The open and challenging issues are also highlighted in this paper.

Underwater Acoustic Sensor Network (UASN) comes with intrinsic constraints because it is deployed in the aquatic environment and uses the acoustic signals to communicate. The examples of those constraints are long propagation delay, very limited bandwidth, high energy cost for transmission, very high signal attenuation, costly deployment and battery replacement, and so forth. Therefore, the routing schemes for UASN must take into account those characteristics to achieve energy fairness, avoid energy holes, and improve the network lifetime. The depth based forwarding schemes in literature use node’s depth information to forward data towards the sink. They minimize the data packet duplication by employing the holding time strategy. However, to avoid void holes in the network, they use two hop node proximity information. In this paper, we propose the Energy and Depth variance-based Opportunistic Void avoidance (EDOVE) scheme to gain energy balancing and void avoidance in the network. EDOVE considers not only the depth parameter, but also the normalized residual energy of the one-hop nodes and the normalized depth variance of the second hop neighbors. Hence, it avoids the void regions as well as balances the network energy and increases the network lifetime. The simulation results show that the EDOVE gains more than 15 % packet delivery ratio, propagates 50 % less copies of data packet, consumes less energy, and has more lifetime than the state of the art forwarding schemes.

Industrial Underwater Acoustic Sensor Networks (IUASNs) come with intrinsic challenges like long propagation delay, small bandwidth, large energy consumption, three-dimensional deployment, and high deployment and battery replacement cost. Any routing strategy proposed for IUASN must take into account these constraints. The vector based forwarding schemes in literature forward data packets to sink using holding time and location information of the sender, forwarder, and sink nodes. Holding time suppresses data broadcasts; however, it fails to keep energy and delay fairness in the network. To achieve this, we propose an Energy Scaled and Expanded Vector-Based Forwarding (ESEVBF) scheme. ESEVBF uses the residual energy of the node to scale and vector pipeline distance ratio to expand the holding time. Resulting scaled and expanded holding time of all forwarding nodes has a significant difference to avoid multiple forwarding, which reduces energy consumption and energy balancing in the network. If a node has a minimum holding time among its neighbors, it shrinks the holding time and quickly forwards the data packets upstream. The performance of ESEVBF is analyzed through in network scenario with and without node mobility to ensure its effectiveness. Simulation results show that ESEVBF has low energy consumption, reduces forwarded data copies, and less end-to-end delay.