The datasets that appear in publications are curated and have been split into training, testing and validation sub-datasets by domain experts. Consequently, machine learning models typically perform well on such split-by-hand prepared datasets. Whereas preparing real-world datasets into curated split training, testing and validation sub-dataset requires extensive effort. Usually, repetitive random splits are carried out and trained and evaluated on until reaching out a good score on the evaluation metrics. In this paper, an algorithmic method is proposed for preparing the sub-datasets splits for machine learning models. The objective of the proposed method is to achieve an evenly representative splits out of the dataset with standard and algorithmic way that reduce the perplexity of random splitting.

Energy efficiency is one of the key aspects of IoT and Wireless Sensor Networks (WSNs) since the nodes of the network are running on battery power and the lifespan of the system is an important mission-critical parameter. In WSNs, where the energy consumption mainly depends on the radio interface and the transmission protocols, reliable packet forwarding from the source node to Base Station (BS) is crucial. In this paper, we focus on developing new routing algorithms which extend the lifespan of WSNs by achieving optimal energy balancing subject to the criterion that the packets must reach the BS with a predefined probability. We will propose novel two-hop and multi-hop routing algorithms to achieve this objective. The performance of the novel algorithms is compared with the LEACH routing protocol. Extensive simulations prove that the new routing methods are indeed energy efficient, and they are able to meet the predefined reliability criteria as well. The results given in this paper can contribute to reliable communication in IoT or WSN networks and result in lower energy consumption.

The Internet of Things (IoT) is transforming the surrounding everyday physical objects into an ecosystem of information that enriches our everyday life. The IoT represents the convergence of advances in miniaturization, wireless connectivity, and increased data storage and is driven by various sensors. Sensors detect and measure changes in position, temperature, light, and many others; furthermore, they are necessary to turn billions of objects into data-generating “things” that can report on their status and often interact with their environment. Application and service development methods and frameworks are required to support the realization of solutions covering data collection, transmission, data processing, analysis, reporting, and advanced querying. This paper introduces the SensorHUB framework that utilizes the state-of-the-art open source technologies and provides a unified tool chain for IoT related application and service development. SensorHUB is both a method and an environment to support IoT related application and service development; furthermore, it supports the data monetization approach, that is, provides a method to define data views on top of different data sources and analyzed data. The framework is available in a Platform as a Service (PaaS) model and has been applied for the vehicle, health, production lines, and smart city domains.

Current video streaming services use a conventional, client-server network topology that puts a heavy load on content servers. Previous work has shown that Peer-to-Peer (P2P) assisted streaming solutions can potentially reduce this load. However, implementing P2P-assisted streaming poses several challenges in modern networks. Users tend to stream videos on the go, using their mobile devices. This mobility makes the network difficult to orchestrate. Furthermore, peers have to contribute their storage to the network, which is challenging, since mobile devices have limited resources compared to desktop machines. In this paper, we introduce an analytical framework for mobile P2P-assisted streaming to estimate the server load that we define as the minimum required server upload rate. Using our framework, we evaluate four caching strategies: infinite cache as a baseline, first in first out (FIFO), random, and Random Linear Network Coded (RLNC) cache. We verify our anal tical results with empirical data that was obtained by carrying out extensive measurements on our working P2P system. Our results show that when employing FIFO, random, and RLNC caching strategies, the server load converges to that of the infinite cache as the cache size increases. With a limit of 5 P2P connections per peer, we show that using the random caching, peers can store 40% fewer packets and still achieve the same benefit as with FIFO caching. When using the RLNC caching, it is enough to store 50% fewer packets to achieve the same benefit.

JSON Web Tokens (JWT) provide a scalable, distributed way of user access control for modern web-based systems. The main advantage of the scheme is that the tokens are valid by themselves – through the use of digital signing – also imply its greatest weakness. Once issued, there is no trivial way to revoke a JWT token. In our work, we present a novel approach for this revocation problem, overcoming some of the problems of currently used solutions. To compare our solution to the established solutions, we also introduce the mathematical framework of comparison, which we ultimately test using real-world measurements.