User segmentation categorises a large and complex user base into manageable similar groups of users. Existing works encounter challenges when dealing with a sparse dataset and finding insights from the generated clusters. This study has two objectives: (1) to identify an optimal clustering model that can handle a sparse dataset and (2) to extract post-clustering insights via a descriptive persona for each cluster. This study deployed clustering models to handle a behavioural user-interaction dataset with a sparsity rate of 85%. The findings revealed that Density-Based Spatial Clustering of Applications with Noise that leveraged on One-hot Encoding and data representation learning via an autoencoder performed best, with a Silhouette score of 0.36. Subsequently, this study enacted techniques and tools such as classification, SHapley Additive exPlanation value, and manual analysis. Classification and SHAP values were used to identify important features that can differentiate clusters created by different clustering models. Specifically, a linear SHAP explainer object was applied to Logistic Regression had been identified to outperformed Random Forest and Light Gradient Boosting Machine, with an accuracy of 97%. A manual analysis of the central tendencies of these relatively more important features within each cluster was performed to create a descriptive persona. The findings revealed four distinctive personas, namely the “Active User,” “COVID-19 Preventer,” “Inactive User,” and “Average Joe.”

The heat transfer and friction factor of ferrofluid passing through a square tube with an oscillating electromagnetic field were investigated experimentally. The impact of electromagnetic rotating direction, flux, and frequency on heat and flow characteristics has been studied. The Brownian motion of particles has been shown to considerably influence the direction, power, and frequency of electromagnetic rotation, resulting in higher heat transfer. The interruption of electromagnetic flow raises the friction factor even further. In addition, a three-layer back propagation network model is built, with input, hidden, and output layers numbered 5, 17, and 2, respectively. This artificial neural network model performed well statistically, with correlation coefficients ranging from 0.99939 to 0.9996 and mean square errors ranging from 0.0106 to 0.0190. The artificial neural network results match the observed data within ±5% and ±10% error ranges for heat and flow characteristics, respectively. Consequently, this machine learning approach might be used to forecast heat exchanger thermal performance.

Video tampering detection is a crucial research area that aims to identify and mitigate the manipulation and tampering of video content. With the widespread use of videos in various applications, ensuring the integrity and truthfulness of video data has become paramount. This research provides a thorough examination of video tampering detection strategies, encompassing both active and passive approaches. The article discusses several different kinds of tampering attacks, such as spatial, temporal, and spatiotemporal tampering, and explores the techniques used to detect these tampering instances. Moreover, the article highlights the challenges and limitations of existing techniques, such as the need for specialized hardware, the vulnerability to owner tampering, and the impact of compression artifacts. The article concludes by highlighting the future directions for research in video tampering detection, including the exploration of deepfake detection and adversarial techniques. The findings of this research provide valuable insights for developing robust and effective video tampering detection systems, thereby enhancing the trustworthiness and reliability of video content in diverse domains.

Accurately predicting the Ultrasonic Pulse Velocity (UPV) in concrete is important invarious fields, including construction, engineering and non-destructive testing. This research paper proposes a novel mathematical model based on Artificial Neural Networks (ANN) applied to accurately predict the UPV of concrete. The proposed model was formulated and validated through comprehensive experimental measurements, which were divided into three subsets for training, validating and testing the model. The training process involved adjusting the model’s parameters through iterative optimization techniques to minimize prediction errors. A separate group of measurements was used to validate the model’s precision and generality. The results demonstrate that the suggested equation matches the data from experiments well. Remarkably, the model achieved a high level of accuracy in predicting UPV values, with an error rate of less than 2% when compared to the measured experimental values. These results validate the effectiveness of the developed ANN model for accurate UPV predictions in concrete structures. The practical applications of the model extend to concrete characterization and non-destructive testing, enabling efficient quality control and assessment of structural integrity. Additionally, the model facilitates concrete characterization and mix design optimization, developing more durable and sustainable structures.

In the present era, increased levels of air pollution have become a major threat to humankind, ecosystems, and climate. Nowadays, the level of noxious emissions in the environment is increasing tremendously owing to industrialization, urbanization, and population growth. The increased pollutant concentrations affect ecosystems, meteorological factors, and human health issues. In light of this, a portable device that can monitor and quantify air quality and harmful gas emissions in response to environmental factors like humidity and temperature is presented in this work. An Internet of Things (IoT) platform with ThingSpeak is employed to monitor all air quality data in real-time by inculcating Air Quality Monitoring System (AQMS). By comparing the recorded data with the standard parameters of the Air Quality Index (0-50 PPM) and NOx emission with 250-350 ppm, the data may be analysed with a data retrieval rate of 2 seconds from the cloud to smart devices. This study will aid in developing an adaptive method for forecasting climatic conditions using a Markov autoregressive model to alarm the environment to take remedial action on the desired environmental aspects.

Owing to the continuous usage of centrifugal pump machines (CPM), health-monitoring systems are important for improving the hydraulic machine industry. While operating the CPM, supervision is carried out by either the operator or the online health-monitoring system. The Health Monitoring System (HMS) helps to understand the conditions of the pump and maintenance prospectus. For breakdown and accidents in the CPM industry, the prime reasons are impeller failure, bearing failure, cavitation, and shaft misalignment; therefore, it is necessary to supervise the parameters to provide a better life to the plant and enhance its efficiency using Data Acquisition (DAQ). The independent features and dependent features are selected and experimentally investigated to examine the characteristics of the CPM plant. The feature selection and extraction process should be compatible with the DAQ system which will help to reduce the curse of data biases and data preparation. Based on that three different DAQ systems based on Arduino, Raspberry Pi, and Dewsoft FFT are used in the experiments and compared based on 5-Point summary, Standard Deviation, cost analysis, flexibility, and maintenance. This paper presents the various techniques and procedures for acquiring the data from CPM using all three DAQ systems and selecting one of them through a statistical approach.

The Box and Block Test (BBT) is a reliable outcome measure in upper limb rehabilitation. The present work aims to automate the test of unilateral manual dexterity of a patient's hand. A box consisting of 80 blocks was divided into two partitions. A motion tracking sensor (MPU6050) was placed on a wrist to obtain the orientation of hand movement. Thirty-four healthy participants were chosen for the test. Each participant completed both traditional and automated versions of BBT within a one-week interval, with each session lasting 60 seconds. A colour sensor at the box monitored block movements across the partitions. A Liquid Crystal Display (LCD) and a 4-digit 7-segment display provided feedback on the number of cubes moved and timing information, respectively. An independent t-test revealed a non-significant difference between traditional BBT and automated BBT. Concurrent validity was established by correlating the results of the automated BBT with the traditional BBT. A positive correlation (r= 0.94) was found between traditional and automated BBT. This validates the automated BBT report with the same result as traditional BBT. Automated BBT can assist a broad range of patients in independently tracking the progress of their hand injuries or conditions without needing a clinician. This can be conveniently done from the comfort of their homes. HighlightsLack of fine motor skills and dexterity is a primary concern affecting the person’s daily living activity; hence, the automated box and block test (BBT) is valid for assessing hand dexterity quantitatively.The traditional BBT is an old and golden test used in dexterity assessment with neurological patients, which only provides outcomes in the form of box count, which is insufficient for hand assessment.The need for this work arises from the limitations of the traditional BBT, which only provides a basic outcome in the form of box counts, insufficient for comprehensive hand dexterity assessment, particularly in neurological patients.Automated BBT is developed, which reduces the clinician’s burden during assessment and makes the test foolproof by adding a mechanism to capture hand trajectory during block movements.Along with the block count, the accelerometer gives hand orientation during the test process, which helps the clinician further analyze the results, such as motion pattern analysis, temporal analysis and comparative analysis.The independent t-test between traditional BBT and automated BBT, p = 0.109, was observed with a 0.95 confidence interval. Levene’s test revealed no significant difference between the scores of the traditional and automated BBT (t = 2.03; p = 0.109), also between dominant and non-dominant hand (t = 1.99; p = 0.74). A positive Pearson’s correlation (r = 0.94, p < 0.05) between traditional and automated tests was obtained.

This study investigates the effect of thermal barrier coatings (TBCs) on the temperature distribution throughout a first-stage high-pressure aero-engine gas turbine rotor blade. The temperature changes on the naked and TBC-coated blade surfaces were investigated using computational models in ANSYS. The results showed a considerable drop in maximum temperature after using yttum-stabilized zirconia (YSZ) as a TBC, demonstrating its efficiency in alleviating thermal stresses. A comparative investigation of various TBC materials revealed varying degrees of heat resistance, with those with lower thermal conductivities. YSZ brought down the maximum temperature of the blade body by 18%. Lanthanum zirconate (La2Zr2O7) showed the best results, bringing down the maximum temperature by 19.5% compared to YSZ. These findings highlight the need to evaluate material qualities other than heat resistance when selecting TBCs for aerospace applications. This study underlines TBCs’ potential to improve the durability and performance of gas turbine rotor blades under high-temperature operating conditions. Furthermore, it suggests future research directions, such as investigations into TBC mechanical properties, erosion resistance, and long-term durability of TBCs, as well as the development of novel TBC formulations and manufacturing techniques to advance aerospace technology and ensure the sustainability of gas turbine operations.

This study investigates the geotechnical factors influencing a recent landslide event in Gera Woreda, Ethiopia. The primary objective was to identify the soil properties contributing to landslide occurrence and understand the triggering mechanisms. Field investigations, soil sampling from both affected and unaffected areas, and subsequent laboratory characterization were conducted. The analysis revealed a dominance of fine-grained soils, such as clay and silt, which are susceptible to weakening upon saturation, thereby increasing landslide susceptibility. Rainfall is identified as the primary trigger for the landslide. Numerical stability assessments using the Limit Equilibrium Method (software, Slide™) and the Finite Element Method (software, PLAXIS®) were performed to assess the stability of the slopes. The stability analysis revealed a notable decrease in the factor of safety (FS) under rising groundwater levels. For example, the FS for Slope 1 decreased from 1.42 under dry conditions to 0.73 at a 2 m groundwater depth. Similarly, Slope 2's FS decreased from 2.06 to 1.18 under similar conditions. These results emphasize that rainfall is the primary trigger for landslides in the area. To address this, the study proposes surface drainage systems and the planting of Vetiver grass to improve slope stability. These findings provide critical insights for developing proactive mitigation strategies to protect local communities and infrastructure from landslide hazards.

Pure and copper-doped zinc oxide (ZnO) nanofibers (NFs) were fabricated by the sol-gel electrospinning (ES) technique with varying concentrations (1, 3, and 5 w%). Scanning electron microscope (SEM), X-ray diffraction (XRD), and ultraviolet-visible (UV-Vis) spectroscopy were used to investigate the morphology, structure, and optical properties of the NFs. The average fiber diameter (AFD) of the Cu-doped ZnO (Cu-Z) NFs decreased with the increase in doping concentration. The structural analysis found that all the NFs samples were polycrystalline and oriented along the (101) plane. Also, the crystalline size decreased from 26.58 to 18.45 nm as the copper (Cu) concentration increased, and a maximum crystalline size was observed for 5 w% Cu-Z NFs. The band gap of undoped and Cu-Z NFs was calculated from UV VIS spectra, and it was observed that the bandgap decreased from 3.45 to 2.9 eV due to the addition of 5 w% Cu. The sensing performance of the NFs was investigated for 50 ppm of ethanol vapor at various operating temperatures (25–300 °C). The results indicated that the 5 w% Cu-Z NFs had the highest response to ethanol vapor, of 12.65 for 50 ppm, with response and recovery times of 70s and 35s. The Cu-Z NFs exhibited good selectivity and stability while tested with other vapors.