- Conference Article
- 10.1109/bigdata66926.2025.11402423
Taxonomy-Based Negative Sampling in Personalized Semantic Search for E-Commerce
- Dec 08, 2025
- Uthman Jinadu + 6 more +6
Publications from 2021 to 2026
Showing 10 of 30 papers
Taxonomy-Based Negative Sampling in Personalized Semantic Search for E-Commerce
Design, Analysis, and Optimization of Satellite Communication Networks Systems for Enhanced Global Connectivity
Satellite Communication Networks System Systems for satellite communication networks include the technological set-ups and supporting infrastructure used for the transmission & reception of voice, video, and data signals. By utilising satellite technology to help with the transmission of information across several Earth locations, these networks enable long-distance communication. Ground stations, satellite systems, & user terminals are the main components of satellite communication networks. The importance of communication via satellite network structures in research cuts across several fields, underscoring the critical role they play. These systems play a crucial role in creating global connectedness and facilitating communication, highlighting the importance of this area of study. Satellite communication networks have a number of advantages, including the ability to connect to remote or inaccessible areas, abundant bandwidth capacity, and global coverage. They have several uses, including in remote sensing, disaster management, internet connectivity, broadcasting, and telecommunications. The Additional Ratio rating approach is a decision-making process used to evaluate and rank options by taking into account a number of factors. It comprises giving each criterion a weight and assessing how well the alternatives perform in comparison to a reference option. The ARAS technique calculates an overall score over each alternative by combining these weighted performance ratios. The choice with the greatest score is considered to be the best one. In situations where several criteria are present, the ARAS technique provides decision-makers with a systematic framework for efficiently evaluating and comparing alternatives. Laser communication, optical networks, satellite optical communication, vibrations, satellite networks Solar Radiation Pressure, Thermal Bending, Solar and Lunar Gravity, Micro Meteorite impacts, Earth oblations effects Through the rank table, we can get the rank of alternative parameters. Whereas satellite optical communication is in 1st position and satellite networks is in 5th position. First ranking satellite optical communication is obtained with the lowest quality of satellite networks.
Read moreBiodiesel production from waste cooking oil (WCO) presents a compelling opportunity to transform discarded oil into a renewable energy resource.Through the conversion of WCO into biodiesel, not only is waste effectively reduced, but a greener, more sustainable alternative to conventional fossil fuels is provided-furthering the shift towards environmentally conscious energy solutions.The importance of this research cannot be overstated.It plays a crucial role in advancing sustainable energy practices, especially by tapping into WCO as a viable and underutilised feedstock for biodiesel production.Consider the scale of global WCO generation: in Canada alone, 135,000 tons are produced annually, while in Asia, the figures soar to a staggering 5.5 million tons.The vast potential for converting this surplus waste into high-value biofuel not only promises substantial environmental benefits but also unlocks significant economic opportunities.The methodology leveraged three distinct machine learning models: Linear Regression (LR), Random Forest Regression (RFR), and Support Vector Regression (SVR).These models were rigorously trained and tested on experimental data derived from biodiesel production processes.The study delved into four critical parameters: Free Fatty Acid (FFA) content, fluctuating between 1.7% and 3.5%, moisture percentage ranging from 0.05% to 0.3%, viscosity measured at 35 to 43 cSt, and reaction time spanning 2 to 3.3 hours.The results were striking, underscoring the robust predictive power of all three models.SVR stood out, achieving the highest training accuracy (R = 0.998), while RFR exhibited a remarkable ability to generalise well on unseen test data (R = 0.989).The analysis uncovered compelling correlations: notably, a robust negative relationship between FFA content and biodiesel yield (-0.91), alongside a positive correlation between viscosity and yield (0.85).These findings underline the capacity of machine learning models to accurately predict biodiesel yields from waste cooking oil (WCO).Each model revealed unique strengths, yet even the simpler Linear Regression model, with an impressive R of 0.979 on test data, pointed to a predominantly linear link between the process parameters and the final yield.Such insights provide invaluable guidance for refining industrial biodiesel production processes, championing the shift towards sustainable energy alternatives and addressing the pressing issues of waste management.
Read moreDeep Multi-Task Learning for Spatio-Temporal Incomplete Qualitative Event Forecasting
Forecasting spatiotemporal social events has significant benefits for society to provide the proper amounts and types of resources to manage catastrophes and any accompanying societal risks. Nevertheless, forecasting event subtypes are far more complex than merely extending binary prediction to cover multiple subtypes because of spatial heterogeneity, experiencing a partial set of event subtypes, subtle discrepancy among different event subtypes, nature of the event subtype, spatial correlation of event subtypes. We present <underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">D</u>e<underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e</u>p mul<underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">t</u>i-task l<underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e</u>arning for spatio-temporal in<underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">c</u>omple<underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">t</u>e qual<underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i</u>tative e<underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">v</u>ent for<underline xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">e</u>casting (DETECTIVE) framework to effectively forecast the subtypes of future events by addressing all these issues. This formulates spatial locations into tasks to handle spatial heterogeneity in event subtypes and learns a joint deep representation of subtypes across tasks. This has the adaptability to be used for different types of problem formulation required by the nature of the events. Furthermore, based on the “first law of geography”, spatially-closed tasks share similar event subtypes or scale patterns so that adjacent tasks can share knowledge effectively. To optimize the non-convex and strongly coupled problem of the proposed model, we also propose algorithms based on the Alternating Direction Method of Multipliers (ADMM). Extensive experiments on real-world datasets demonstrate the model’s usefulness and efficiency.
Read moreEnhancing Autonomous Vehicle Perception: A 6-DOF Pose Estimation Framework Using Deep Learning
This research addresses the critical challenge of precise perception of surrounding vehicles' positions and orientations in autonomous driving environments. We aim to develop an algorithm to estimate the absolute 6-degree-of-freedom (6-DOF) pose of vehicles from single images using a dataset of over 60,000 labeled 3D vehicle instances. Our approach integrates Faster R-CNN for 2D bounding box detection, Q-Net for 6D-pose regression, and DenseNet for enhanced feature extraction. Data analysis revealed attribution errors in pitch and yaw and significant variance in spatial accuracy across different vehicle models. By fine-tuning the Faster R-CNN on BDD100k and our dataset, and employing weighted bounding box and point fusion techniques, our model outperforms existing methods in accuracy and robustness. This work provides a reliable method for autonomous vehicles to perceive their surroundings accurately, enhancing navigation safety and efficiency.
Read moreKnowledge Graph-based Session Recommendation with Session-Adaptive Propagation
Session-based recommender systems (SBRSs) predict users' next interacted items based on their historical activities. While most SBRSs capture purchasing intentions locally within each session, capturing items' global information across different sessions is crucial in characterizing their general properties. Previous works capture this cross-session information by constructing graphs and incorporating neighbor information. However, this incorporation cannot vary adaptively according to the unique intention of each session, and the constructed graphs consist of only one type of user-item interaction. To address these limitations, we propose knowledge graph-based session recommendation with session-adaptive propagation. Specifically, we build a knowledge graph by connecting items with multi-typed edges to characterize various user-item interactions. Then, we adaptively aggregate items' neighbor information considering user intention within the learned session. Experimental results demonstrate that equipping our constructed knowledge graph and session-adaptive propagation enhances session recommendation backbones by 10%-20%. Moreover, we provide an industrial case study showing our proposed framework achieves 2% performance boost over an existing well-deployed model at The Home Depot e-platform.
Read moreMachine Learning-Driven Reliability Engineering for E-Commerce Sites: A Study on Browsing, Cart, and Checkout Phases
This study explores reliability engineering in e-commerce systems by analyzing critical user journey stages – browsing, cart, and checkout – using advanced machine learning models. The research uses random forest regression (RFR) and support vector regression (SVR) to assess and predict system performance, stability, and potential failure points. Data was obtained from simulated e-commerce transactions that included key parameters such as response time, server load, transaction success rate, and error frequency. The models were evaluated using statistical indicators including coefficient of determination (R²), mean square error (MSE), and mean absolute error (MAE) to measure predictive accuracy. The results indicated that both models achieved high reliability and accuracy, with SVR showing slightly higher generalization ability. The findings highlight that predictive modeling can effectively identify performance bottlenecks, reduce downtime, and improve the overall robustness of e-commerce operations. By integrating reliability engineering principles with machine learning techniques, this study contributes to improving user experience, enhancing system resilience, and ensuring scalable and stable online transaction environments. This hybrid analytics approach provides a foundation for developing proactive monitoring strategies in modern e-commerce infrastructures. Keywords: Reliability Engineering, E-Commerce, Random Forest Regression, Support Vector Regression, Performance Optimization Objective: Improve the reliability and efficiency of the browsing, cart, and checkout processes, ensuring a seamless and consistent e-commerce user experience
Read moreSafety Study of an Original Mesenchymal Stromal Cell Secretome-Based Medicinal Product for Spermatogenesis Restoration
SCIENTIFIC RELEVANCE. Currently, there are no effective and safe medicinal products for idiopathic male infertility. Previous studies in two animal models of infertility (short-term cryptorchidism in rats and doxorubicin-induced testicular injury in mice) have shown the effectiveness of an originator medicinal product based on the mesenchymal stromal cell (MSC) secretome.AIM. The aim of the study was to evaluate the toxicity profile of the MSC secretome-based medicinal product in rats after local intratesticular or intramuscular administration.MATERIALS AND METHODS. The MSC secretome is a combination of factors secreted by MSCs in low-glucose Dulbecco’s modified Eagle’s medium (DMEM-LG) for MSC conditioning. In the single-dose toxicity study, the MSC secretome-based medicinal product was injected under the testicular tunica albuginea of male Wistar rats (15 per group) at doses of 15 and 25 relative units (RU) per animal, which are 1.5 and 2.5 times higher than the therapeutic dose (10 RU). In the repeat-dose toxicity study, male Wistar rats (10 per group) received intramuscular thigh injections of the medicinal product on days 1, 6, and 12 at doses of 15 and 25 RU per animal. The local tolerance study involved histopathological examination of the testes and thighs at the injection site. All studies included control groups of intact animals and animals similarly injected with blank DMEM-LG. The early follow-up period was 14 days, and the late follow-up period was 42 days.RESULTS. The rats showed no changes in the general condition after single and repeated doses of the MSC secretome-based medicinal product. Single subtunical doses induced moderate irritation; its signs included pathological changes in individual seminiferous tubules: epithelial atrophy (70% of the animals on day 14; 55% at late follow-up) and sperm stasis (70% of the animals). Similar changes were observed in the blank DMEM-LG group (up to 80% of the animals). There were no pathological changes in the tissues after repeated injections. A transient increase in alkaline phosphatase activity was detected in animals after their third intramuscular injection at a dose of 25 RU; the other biochemical parameters were normal in all study groups.CONCLUSIONS. The MSC secretome-based medicinal product has a favourable safety profile following both intratesticular and intramuscular administration, as it does not cause any permanent changes in the studied organs and tissues.
Read moreComparative Preclinical Evaluation of the Safety, Antifungal Activity, and Pharmacokinetics of Sertaconazole Products for External Use
SCIENTIFIC RELEVANCE. The high prevalence of fungal skin infections motivates expanding the range of sertaconazole products for external use.AIM. The study was a preclinical comparison of the safety, antifungal activity, and pharmacokinetics of Sertaverin® 2% medicated shampoo (VERTEX JSC, Russia) with those of Sertamicol® 2% solution for external use (Glenmark Pharmaceuticals Ltd, India) and Nizoral® 2% shampoo (Janssen Pharmaceuticals N.V., Belgium) approved in the Russian Federation.MATERIALS AND METHODS. In the toxicity study, the medicinal products were applied to the skin of male and female outbred rats at doses of 0.5 or 1.5 mL/animal for 28 days. The authors evaluated the pharmacokinetics of two sertaconazole formulations (shampoo and solution) following a single administration to adult male rats at the same dose. Nizoral® was not used in the pharmacokinetics study because it contains a different active substance, ketoconazole. The minimum inhibitory concentration (MIC) was determined using the serial microdilution method in a wide range of concentrations.RESULTS. The medicinal products did not exhibit any significant toxic effects in laboratory animals after 28 days of repeated dermal application. Plasma sertaconazole concentrations were negligible. Sertaconazole was intensively distributed in the liver, which is a highly vascularised organ, and in the target organ (skin at the site of application). The relative bioavailability of sertaconazole from the shampoo relative to that from the solution for external use was approximately 30% in liver tissues and approximately 363% in skin tissues at the application site. Sertaverin® was comparable to sertaconazole in the active substance form in terms of inhibiting the growth of Malassezia furfur strains. The MICs calculated on the active substance basis were ≤16–64 μg/mL.CONCLUSIONS. With its synergistic dual mechanism of action, broad-spectrum antifungal activity, lipophilic properties, and low systemic absorption, Sertaverin® may provide a more effective and safe alternative to marketed medicinal products for scalp diseases.
Read moreA Study of Embryotoxic, Foetotoxic, and Teratogenic Effects of the Original Antituberculosis Agent Thiozonide in Pregnant Rabbits
Several treatment regimens with antituberculosis medicinal products are available for tuberculosis. Thiozonide is a newly developed original antimicrobial agent that exhibits bacteriostatic activity against Mycobacterium tuberculosis strains H37Rv, CN-37, CN-40, and MS-115.The aim of the study was to investigate the embryotoxic, foetotoxic, and teratogenic effects of thiozonide in pregnant rabbits.Materials and methods. The study involved 66 pregnant rabbits (4 groups of 16–17 animals each). The rabbits received oral thiozonide from day 6 to day 19 of gestation at doses of 20.6 mg/kg (1 TD (therapeutic dose)), 103 mg/kg (5 TDs), and 206 mg/kg (10 TDs). The control group received a 1% starch solution. The authors conducted a macroscopic examination of the reproductive organs and a histological evaluation of the placenta in eutha nised pregnant rabbits. Live foetuses underwent a check for developmental abnormalities, a skeletal development evaluation with a modified Dawson’s method, and a histologic examination of the internal organs.Results. The study showed no clinical signs of toxicity and no mortality associated with thiozonide in pregnant rabbits across all dose groups. Macroscopic and histological examinations revealed no pathological changes in the reproductive organs of pregnant rabbits. The evaluation of embryotoxic and foetotoxic effects did not identify any differences between the foetuses of the animals assigned to different doses of thiozonide and the control group. The authors found no developmental abnormalities in the foetuses. Examinations of foetal skeleton development and internal organ condition identified no differences between the groups and no abnormalities. The authors registered the death of all foetuses (late resorption) in one rabbit from the 206 mg/kg group. Therefore, the 103 mg/kg dose (5 TDs) was selected as a reasonable No Observed Adverse Effect Level (NOAEL).Conclusions. Thiozonide has no embryotoxic, foetotoxic, or teratogenic effects.
Read more