Data-driven Strategy Research Articles

Designing of data-driven strategies for the online one-way trading problem

This paper aims to propose a novel framework for designing data-driven strategies for online problems, which utilizes historical data to enhance the actual execution effect of online strategies. It integrates the concepts of competitive ratio and competitive difference to form a comprehensive competitive power concept. Using this framework, a threat-based strategies, TBSλ (threat-based strategy with preference λ) is proposed as an optimal strategy in the sense of competitive power for the one-way trading problem, and a data-driven online strategy DDOS (data-driven online strategy) is further proposed. By the discussion of special cases, it is found that TBS0, which is based on competitive difference, is quite different from CDA (competitive difference analysis) in former literature but is also an optimal strategy for the one-way trading in the sense of competitive difference. It is further found that strategy TBS0 performs more aggressively than strategy TBS1 which is based on competitive ratio through special case analysis. Additionally, numerical experiments based on China’s carbon emission trading markets are presented to further validate the proposed online strategy. The results demonstrate the superiority of DDOS over the ones based solely on competitive ratio or competitive difference.

Balanced Quality Score: Measuring Popularity Debiasing in Recommendation

Popularity bias is the tendency of recommender systems to further suggest popular items while disregarding niche ones, hence giving no chance for items with low popularity to emerge. Although the literature is rich in debiasing techniques, it still lacks quality measures that effectively enable their analyses and comparisons. In this article, we first introduce a formal, data-driven, and parameter-free strategy for classifying items into low, medium, and high popularity categories. Then we introduce Balanced Quality Score (BQS) , a quality measure that rewards the debiasing techniques that successfully push a recommender system to suggest niche items, without losing points in its predictive capability in terms of global accuracy. We conduct tests of BQS on three distinct baseline collaborative filtering frameworks: one based on history-embedding and two on user/item-embedding modeling. These evaluations are performed on multiple benchmark datasets and against various state-of-the-art competitors, demonstrating the effectiveness of BQS.

Data-Driven Financial Strategy Transformation: The Impact of Corporate Data Capital Utilization on The Level of Financial Asset Allocation

Data-Driven Financial Strategy Transformation: The Impact of Corporate Data Capital Utilization on The Level of Financial Asset Allocation

A Robust Sparse Sensor Placement Strategy Based on Indicators of Noise for Ocean Monitoring

A well-performing data-driven sparse sensor deployment strategy is critical for marine monitoring systems, as it enables the optimal reconstruction of marine physical quantities with fewer sensors. However, ocean data typically contain substantial amounts of noise, including outliers (incomplete data) and inherent measurement noise, which heightens the complexity of sensor deployment. Therefore, this study optimizes the sparse sensor placement model by establishing noise indicators, including small noise weight and large noise weight, which are measured by entropy to minimize the prediction bias. Building on this, a robust sparse sensor placement algorithm is proposed, which utilizes the block coordinate update (BCU) iteration method to solve the function. During the iterative updating process, the proposed algorithm simultaneously updates the selection matrix, reconstruction matrix, and noise matrix. This allows for effective identification and mitigation of noise in the data through evaluation. Consequently, the deployed sensors achieve superior reconstruction performance compared to other deployment methods that do not incorporate noise evaluation. Experiments are also conducted on datasets of sea surface temperature (SST) and global ocean salinity, which demonstrate that our strategy significantly outperforms several other considered methods in terms of reconstruction accuracy while enabling autonomous sensor deployment under noisy conditions.

Unlocking the AI-powered customer experience: Personalized service, enhanced engagement, and data-driven strategies for e-commerce applications

In recent times, there has been a surge of interest in the transformative potential of artificial intelligence (AI), particularly within the realm of online advertising. This research focuses on the critical examination of AI’s role in enhancing customer experience (CX) across diverse business applications. The aim is to identify key themes, assess the impact of AI-powered CX initiatives, and highlight directions for future research. Employing a systematic and comprehensive approach, the study analyzes academic publications, industry reports, and case studies to extract theoretical frameworks, empirical findings, and practical insights. The findings underscore a significant transformation catalyzed by AI integration into Customer Relationship Management (CRM). AI enables personalized interactions, fortifies customer engagement through interactive agents, provides data-driven insights, and empowers informed decision-making throughout the customer journey. Four central themes emerge: personalized service, enhanced engagement, data-driven strategy, and intelligent decision-making. However, challenges such as data privacy concerns, ethical considerations, and potential negative experiences with poorly implemented AI persist. This article contributes significantly to the discourse on AI in CRM by synthesizing the current state, exploring key themes, and suggesting research avenues. It advocates for responsible AI implementation, emphasizing ethical considerations and guiding organizations in navigating opportunities and challenges.

Unraveling Reactivity Origin of Oxygen Reduction at High-Entropy Alloy Electrocatalysts with a Computational and Data-Driven Approach.

High-entropy alloys (HEAs), characterized as compositionally complex solid solutions with five or more metal elements, have emerged as a novel class of catalytic materials with unique attributes. Because of the remarkable diversity of multielement sites or site ensembles stabilized by configurational entropy, human exploration of the multidimensional design space of HEAs presents a formidable challenge, necessitating an efficient, computational and data-driven strategy over traditional trial-and-error experimentation or physics-based modeling. Leveraging deep learning interatomic potentials for large-scale molecular simulations and pretrained machine learning models of surface reactivity, our approach effectively rationalizes the enhanced activity of a previously synthesized PdCuPtNiCo HEA nanoparticle system for electrochemical oxygen reduction, as corroborated by experimental observations. We contend that this framework deepens our fundamental understanding of the surface reactivity of high-entropy materials and fosters the accelerated development and synthesis of monodisperse HEA nanoparticles as a versatile material platform for catalyzing sustainable chemical and energy transformations.

Efficient Identification Method for Power Quality Disturbance: A Hybrid Data-Driven Strategy

Efficient Identification Method for Power Quality Disturbance: A Hybrid Data-Driven Strategy

Convolutional Encoding and Normalizing Flows: A Deep Learning Approach for Offshore Wind Speed Probabilistic Forecasting in the Mediterranean Sea

Abstract The safe and efficient execution of offshore operations requires short-term (1–6 h ahead) high-quality probabilistic forecasts of metocean variables. The development areas for offshore wind projects, potentially in high depths, make it difficult to gather measurement data. This paper explores the use of deep learning for wind speed forecasting at an unobserved offshore location. The proposed convolutional architecture jointly exploits coastal measurements and numerical weather predictions to emulate multivariate probabilistic short-term forecasts. We explore both Gaussian and non-Gaussian neural representations using normalizing flows. We benchmark these approaches with respect to state-of-the-art data-driven schemes, including analog methods and quantile forecasting. The performance of the models and resulting forecast quality are analyzed in terms of probabilistic calibration, probabilistic and deterministic metrics, and as a function of weather situations. We report numerical experiments for a real case study off the French Mediterranean coast. Our results highlight the role of regional numerical weather prediction and coastal in situ measurement in the performance of postprocessing. For single-valued forecasts, a 40% decrease in RMSE is observed compared to the direct use of numerical weather predictions. Significant skill improvements are also obtained for the probabilistic forecasts, in terms of various scores, as well as an acceptable probabilistic calibration. The proposed architecture can process a large amount of heterogeneous input data and offers a versatile probabilistic framework for multivariate forecasting.

Damage Identification at Off-Grid Locations on Plate-like Structures using Frequency-Coupled Group Lasso

Localizing elastic wave scatterers is the working principle behind a family of model-based approaches to detect and localize damage in plate-like structures. These approaches take advantage of a physical model of propagation and scattering, which enables an efficient data-driven strategy to decompose the effect of damage on the structure. This strategy consists in measuring on the healthy structure the responses to excitations at predefined candidate damage locations, and then in building from these a dictionary that maps damage at candidate locations to changes in responses. At inspection time, inverse methods are used to identify damage, based on this dictionary and the responses measured on the structure under test. One such inverse method is Frequency-Coupled Group Lasso, which exploits the inherent sparsity of damage to reduce the number of required measurements. This paper examines the performance of the above-described approach in the “off-grid” case, that is, when the damage is not located at one of the predefined candidate locations. This is an important extension because real-world damage can occur at a large or infinite number of locations, but the number of candidate locations is limited by measurement and/or computing cost. Off-grid cases are simulated by removing from the dictionary the entries corresponding to the actual damage locations and then running Frequency-Coupled Group Lasso with this incomplete dictionary. The results are then compared to the on-grid case, when the actual damage locations are included in the dictionary. Damage identification is demonstrated on a 600mm by 600mm composite plate with up to 6 damage locations, using a single accelerometer and 7 hammer excitations. Off-grid performance is evaluated for 4 different accelerometer locations, for SNRs ranging from 30 to 10dB, and compared to on-grid performance. The proposed damage identification approach is shown to remain robust in the off-grid case, further proving its real-world applicability.

Ligand control of regioselectivity in palladium-catalyzed heteroannulation reactions of 1,3-Dienes

Olefin carbofunctionalization reactions are indispensable tools for constructing diverse, functionalized scaffolds from simple starting materials. However, achieving precise control over regioselectivity in intermolecular reactions remains a formidable challenge. Here, we demonstrate that using PAd2nBu as a ligand enables regioselective heteroannulation of o-bromoanilines with branched 1,3-dienes through ligand control. This approach provides regiodivergent access to 3-substituted indolines, showcasing excellent regioselectivity and reactivity across a range of functionalized substrates. To gain further insights into the origin of selectivity control, we employ a data-driven strategy, developing a linear regression model using calculated parameters for phosphorus ligands. This model identifies four key parameters governing regioselectivity in this transformation, paving the way for future methodology development. Additionally, density functional theory calculations elucidate key selectivity-determining transition structures along the reaction pathway, corroborating our experimental observations and establishing a solid foundation for future advancements in regioselective olefin difunctionalization reactions.

A data-driven short video international communication model based on indicator system communication network and attention BiLSTM neural network

As one of the most popular ways to disseminate Internet content in the current era of media convergence, short videos have increasingly become part of people’s daily lives. This paper takes TikTok as the research object, focusing on the study of short video communication process. The research objective of this paper is to improve the international communication effect of short videos through data-driven analysis and strategy research on the law of international communication of short videos. Firstly, the key objective indicator system about international communication of short videos is constructed, the definition, selection and design logic of key objective indicators are introduced, the construction idea of the corresponding indicator system is proposed, and then the key objective indicator system about international communication of short videos and related data set are constructed. Secondly, an international communication network model for short videos is built. The construction of this network model is divided into three steps, the construction of the short video communication network KOIS-PN based on the key objective indicators system, the construction of AT-BiLSTM neural network model to adapt to the research scenario of short video international communication, the construction of an international communication network model of short videos combining KOIS-PN and AT-BiLSTM. The international short video communication network is constructed according to the key objective index system of international short video, and the key objective index system is used as nodes in the communication network, and the short video communication network is constructed by hierarchical relationship. Thirdly, the performance of the core theoretical model proposed in this paper is verified. The experimental results show that, the KOISPN-ATBiLSTM international communication network model of short videos has certain validity and advancement.

Prediction of Axial Capacity of RC Columns Reinforced with Ferro-cement Jacketing: A Data-driven Machine Learning Strategy

Prediction of Axial Capacity of RC Columns Reinforced with Ferro-cement Jacketing: A Data-driven Machine Learning Strategy

Heat generation rate estimation of lithium-ion batteries for electric vehicles by BP-based optimized neural network

Accurate estimation of heat generation rate (HGR) of lithium-ion batteries (LIBs) is a critical and essential task for their decent thermal management, thereby facilitating the safe driving of electric vehicles (EVs). In order to improve the accuracy of HGR estimation and reduce the structural complexity of network, a data-driven strategy is developed through integrating Bayesian optimization (BO), Adam optimization, and Principal Component Analysis (PCA) with Back Propagation (BP) neural network. The BO algorithm is utilized to optimize the hyperparameters of the BP neural network for prediction accuracy enhancement. The PCA is employed to extract the feature matrix thereby reducing the complexity of inputs. The Adam optimization algorithm is used to improve computational efficiency. The performance of the proposed strategy was validated based on a dataset derived from lab-scale experiments, as well as a publicly available dataset regarding practical driving conditions. The test results show that the proposed strategy can achieve accurate HGR estimation and result in a mean absolute error (MAE) of 0.0504 W, and a root mean square error (RMSE) of 0.0628 W, and a R2 of 0.9998. Compared to some other HGR estimation methods, the proposed strategy achieved a significant enhancement in accuracy indexes, indicating its superior accuracy and robustness.

Navigating the Skincare Journey: Analysing the Effectiveness of COSRX’s Global Marketing Strategy

Cosmeceuticals are a brand-new product category that has emerged between pharmaceuticals and cosmetics to improve skin’s appearance and health. In the fiercely competitive global skincare market, brands increasingly struggle to navigate the complex customer journey and secure lasting brand loyalty. This research delves into the success story of COSRX, a South Korean skincare brand that carved a distinct niche by prioritising key objectives like revitalising local health and beauty stores and fostering customer satisfaction through a deeply engaged, data-driven digital marketing strategy. By examining these objectives through four key pillars––revitalising local partnerships, customer-centric approach, leveraging Amazon reviews and designing a compelling digital strategy––this paper offers valuable insights and actionable takeaways for brands navigating the ever-evolving landscape of skincare marketing.

Impact of Ramadan on Halal Food Marketing Strategies in the Chinese Market: A Data Analytics Approach

This study investigates the integration of halal food into Chinese multinational culture amid the Coronavirus epidemic. Applying a thorough data analytics procedure, the research focuses on the evolving e-commerce landscape within China. The adaptation under scrutiny includes a wide range of sociocultural transformations, including linguistic proficiency, social norms, dietary preferences, and e-commerce consumption habits from various cultural domains. A data-driven marketing strategy's analytical structure, supported by an updated Segmentation, Targeting, and Positioning (STP) model, carefully examines product differentiation criteria. Furthermore, the investigation includes the implementation of a prediction model that uses complex algorithms to estimate market trends and consumer preferences. The term "halal" comes from Arabic and refers to food that is considered permissible for consumption excluding porcine products, carrion, blood, and alcohol. Halal cuisine, which adheres to Islamic dietary regulations, is critical to Muslims' religious devotion around the globe. Halal dietary habits are integrated into Chinese cuisine through major e-commerce platforms. This study utilizes a data-analytics marketing model to analyze transactional data from Taobao.com and Tmall.com, aiming to understand purchasing patterns of halal food products in China. The analysis highlights significant seasonal fluctuations, particularly during major festivals such as Ramadan, Chinese New Year, and Eid, which heavily influence sales.

Consistent spectral approximation of Koopman operators using resolvent compactification

Koopman operators and transfer operators represent dynamical systems through their induced linear action on vector spaces of observables, enabling the use of operator-theoretic techniques to analyze nonlinear dynamics in state space. The extraction of approximate Koopman or transfer operator eigenfunctions (and the associated eigenvalues) from an unknown system is nontrivial, particularly if the system has mixed or continuous spectrum. In this paper, we describe a spectrally accurate approach to approximate the Koopman operator on L 2 for measure-preserving, continuous-time systems via a ‘compactification’ of the resolvent of the generator. This approach employs kernel integral operators to approximate the skew-adjoint Koopman generator by a family of skew-adjoint operators with compact resolvent, whose spectral measures converge in a suitable asymptotic limit, and whose eigenfunctions are approximately periodic. Moreover, we develop a data-driven formulation of our approach, utilizing data sampled on dynamical trajectories and associated dictionaries of kernel eigenfunctions for operator approximation. The data-driven scheme is shown to converge in the limit of large training data under natural assumptions on the dynamical system and observation modality. We explore applications of this technique to dynamical systems on tori with pure point spectra and the Lorenz 63 system as an example with mixing dynamics.

A Self-Assessment Tool for Helping Identify Police Burnout Among Investigators of Child Sexual Abuse Material

IntroductionLaw enforcement professionals who investigate child sexual abuse material face increased risk of mental health challenges, including burnout. This study aims to develop a data-driven self-assessment tool for law enforcement personnel exposed to child sexual abuse material. The tool assesses burnout symptoms and related mental health issues, offering a proactive approach to identifying and supporting individuals at risk. MethodsA mixed-methods investigation involved 500 police investigators and forensic examiners from the United States. The study utilized a convenience sample recruited through various channels connected with the National Criminal Justice Training Center. ResultsTwenty percent of participants exhibited high burnout. The BURNT demonstrated a sensitivity of 69.6% and specificity of 74.6% at a cutoff point of ≥2; correctly classifying 73.6% of the sample. Individuals with scores of 2 or more were 3.47 times more likely to be experiencing high burnout compared to peers with a score of zero with increasing odds with each additional score. High burnout was associated with longer tenure in current positions. ConclusionsThe BURNT offers a short and simple self-assessment tool for law enforcement professionals exposed to child sexual abuse material, aiding in the early identification of burnout symptoms. A cutoff point of ≥2 provides a data-driven strategy for identifying individuals at increased risk, promoting timely intervention and support to mitigate burnout's adverse effects on mental well-being and professional performance. The BURNT's sensitivity and specificity balance enhances its utility, providing a proactive approach to address the unique mental health challenges faced by law enforcement personnel combating child sexual abuse material.

Experimental validation of data-driven reactive control strategy for wave energy converters: A Gaussian process regression approach

Wave energy is a renewable energy source that can generate electricity without emitting carbon dioxide using wave energy converters (WECs). WECs face several challenges, such as effectively harnessing wave energy, ensuring operational safety, and implementing cost-saving measures. In this study, a data-driven reactive control strategy for WECs using Gaussian process regression is experimentally validated. The control strategy aimed to maximize energy absorption while satisfying physical constraints without a WEC dynamic model. In this study, the Gaussian process regression model represents a WEC dynamic model using an approximation function based on the input/output data. Experimental validation results confirmed that the proposed data-driven reactive control strategy effectively controlled the WEC and improved its performance using the Gaussian model. Moreover, performance was improved under different wave conditions compared to those used during training. The control strategy can be useful in situations where conventional methods cannot realize ideal control parameters, such as when substantial model errors exist or the WEC dynamic model changes over time. The data-driven approach that uses Gaussian process regression reduces learning costs compared to other machine-learning methods. It also has stable control performance because it is independent of the training dataset. The potential of data-driven approaches for optimizing WEC control strategies was highlighted.

An Online Home Energy Management System using Q-Learning and Deep Q-Learning

The users of home energy management systems schedule their real-time energy consumption thanks to advancements in communication technology and smart metering infrastructures. In this paper, a data-driven strategy is proposed, which is an Online Home Energy Management System (ON-HEM) that uses reinforcement learning algorithms (Q-Learning and Deep Q-Learning) to control the optimal energy consumption of a smart home system. The proposed system comprises power resources (grid, photovoltaic), communication networks, and appliances with their agents classified into four groups: deferrable, non-deferrable, power level controllable, and electric vehicle. The system reduces electricity costs and high peak demands while considering the cost of user dissatisfaction with real-life data. Simulations are performed on the proposed ON-HEM considering different pricing approaches (Real Time Pricing and Time of Use Pricing) with Q-Learning and Deep Q-Learning (DQL) algorithms using PyCharm Professional Edition software. The findings demonstrate both the superiority of DQL over Q-Learning and the efficiency of the proposed ON-HEM in decreasing high peak demand, electricity costs, and customer dissatisfaction costs. The efficiency and dependability of the proposed system were verified by utilizing simulation-based findings with real-life data using IBM SPSS Statistics software.

Investigating the Physical and Operational Characteristics of Manufacturing Processes for MFI-Type Zeolite Membranes for Ethanol/Water Separation via Principal Component Analysis

In this study, Principal Component Analysis (PCA) was applied to discern the underlying trends for 31 distinct MFI (Mobil No. 5)-zeolite membranes of 11 textural, chemical, and operational factors related to manufacturing processes. Initially, a comprehensive PCA approach was employed for the entire dataset, revealing a moderate influence of the first two principal components (PCs), which collectively accounted for around 38% of the variance. Membrane samples exhibited close proximity, which prevented the formation of any clusters. To address this limitation, a subset acquisition strategy was followed, based on the findings of the PCA for the entire dataset. This resulted in an enhanced overall contribution and the revelation of diverse patterns among the membranes and the considered manufacturing factors (total variance between 55% and 77%). The segmentation of the data unveiled a robust correlation between silica (SiO2) concentration and pervaporation conditions. Additionally, a notable clustering of the chemical compositions of the preparation solutions underscored their significant influence on the operational efficacy of MFI zeolite membranes. On the other hand, an exclusive chemical composition of the preparation solution was noticed. This highlighted the high influence of the chemical composition on the operational efficiency of MFI zeolite membranes. The coupling of PCA with experimental results can provide a data-driven enhancement strategy for the manufacturing of MFI-type zeolite membranes used for ethanol/water separation.