Multivariate Data Research Articles

REDEFINING CONSUMER ENGAGEMENT: THE IMPACT OF AI AND MACHINE LEARNING ON MARKETING STRATEGIES IN TOURISM AND HOSPITALITY

This article aims to systematically investigate and elucidate the transformative effects of Artificial Intelligence (AI) and Machine Learning (ML) on marketing strategies and consumer engagement within the tourism and hospitality industries. The research methodology employed in this article encompasses a quantitative approach, underpinned by the use of cluster analysis to categorize and interpret complex multivariate data. This methodological framework is chosen to provide a rigorous, data-driven examination of the impacts of Artificial Intelligence (AI) and Machine Learning (ML) on marketing strategies and consumer behavior in the tourism and hospitality industry. The results of this investigation reveal significan t insights into the application of Artificial Intelligence (AI) and Machine Learning (ML) in enhancing marketing strategies within the tourism and hospitality sectors. Through the application of the k-means clustering algorithm to the collected dataset, distinct patterns of consumer behavior and preferences have emerged, underscoring the potential of AI and ML to revolutionize marketing approaches and consumer engagement. The findings from this research underscore the pivotal role of Artificial Intelligence (AI) and Machine Learning (ML) in transforming marketing strategies and enhancing consumer engagement within the tourism and hospitality industries. The utilization of cluster analysis, specifically the k -means algorithm, has facilitated a deeper understanding of consumer behavior patterns, leading to several key conclusions.

Spatio-temporal plant hormonomics: from tissue to subcellular resolution

Abstract Due to technological advances in mass spectrometry, significant progress has been achieved recently in plant hormone research. Nowadays, plant hormonomics is well established as a fully integrated scientific field focused on the analysis of phytohormones, mainly on their isolation, identification, and spatiotemporal quantification in plants. This review represents a comprehensive meta-study of the advances in the phytohormone analysis by mass spectrometry over the past decade. To address current trends and future perspectives, Web of Science data were systematically collected and key features such as mass spectrometry-based analyses were evaluated using multivariate data analysis methods. Our findings showed that plant hormonomics is currently divided into targeted and untargeted approaches. Both aim to miniaturize the sample, allowing high-resolution quantification to be covered in plant organs as well as subcellular compartments. Therefore, we can study plant hormone biosynthesis, metabolism, and signalling at a spatio-temporal resolution. Moreover, this trend has recently been accelerated by technological advances such as fluorescence-activated cell sorting or mass spectrometry imaging.

Self-multivariate spectral decomposition and mode-fused envelope spectrum for enhancing bearing fault feature

In practical diagnostic scenarios, multivariate vibration data (MVD) actually contain more comprehensive fault information. Variational mode extraction (VME) has recently become a promising tool for extracting a specific mode from vibration data. However, its performance of feature extraction is restricted by presetting initial parameters and the joint information within MVD is ignored to be analyzed by VME. To address these issues, a self-multivariate spectral decomposition (SMSD), which includes multivariate spectral feature detector and multivariate mode extraction, is presented to efficiently decouple the fault-related features from MVD. The proposed method can avoid the omitting of effective modes and inclusion of interferential modes. Furthermore, a mode-fused envelope spectrum (MFES) is built to suppress in-band noise of extracted modes. Three cases demonstrate that SMSD can identify multivariate fault-related modes effectively and MFES can exhibit the fault-related frequencies clearly. Moreover, its superior performance is verified by comparing with some well-known multivariate decomposition methods.

Comparative analysis of the aromatic profile of single malt whiskies from different regions of Scotland using GC-MS, GC-O-MS and sensory evaluation

Single malt whiskies are popular around the world. The 37 esters, 10 alcohols, 10 phenols, 8 aldehydes, 8 acids, 6 alkanes, 4 ketones and 3 heterocycles were detected in the single malt whiskies extracted from different origins in Scotland by HS-SPME (headspace solid-phase microextraction) and LLE (liquid-liquid extraction) and identified by GC-MS (gas chromatography-mass spectrometry). QDA (quantitative descriptive analysis) combined with multivariate statistics were used to investigate the effects of the differential aroma-active compounds on the sensory aroma characteristics. O-cresol, ethyl 2-methylbutyrate, 5-methylfurfural and propionic acid might influence the flavor of single malt whiskies of different origins in Scotland. The 54 key whisky aroma compounds were analyzed, with isoamyl acetate having the highest OAV (odour activity value) of all whisky samples. In addition, the aromas of volatile compounds were analyzed using OSME (odor-specific magnitude estimation) and OAV. Based on these facts, the researchers developed a flavour wheel with 10 aroma profiles. This paper evaluates Scotch single malt whisky from a Chinese perspective, which helps the development of the Chinese whisky market.

Distribution of Elements in Durum Wheat Seed and Milling Products: Discrimination between Cultivation Methods through Multivariate Data Analysis.

Many staple foods originate from durum wheat and its milling products; because of this, it is very important to know their characteristics. This study investigates elemental contents in these products and if differences exist because of organic farming. The concentrations of 28 elements in the whole seed and in milling products, that is, bran, semolina and flour, of durum wheat, were determined through ICP-OES. The wheats were grown under conventional or organic agronomic practices to verify the possibility of discriminating, using the elemental content, between products coming from one or the other practice. The elements were more abundant in the outer layer of the seed, the bran, but most of them were also present in the others. Traces of Sb were present only in 3% of the samples, while traces of Tl were detected in approximately half of the seed and bran samples but not in other samples. The absence of an element was more characteristic of specific products, e.g., most semolina and flour lacked Co, while other elements showed small differences between products from organic and conventional cultivation or between different milling products, which was the case, for example, for traces of Ag, B, and V. The concentrations of these elements were coupled with multivariate discriminant analysis, specifically PLS-DA, to identify the cultivation provenance of the milled products. A few elements, although different for each product, are sufficient to attain precision and accuracy of classification close to 1; small differences exist for different products. The worst is flour, where the predicted precision and accuracy are 0.92, although using only three elements: B, K, and Se. Semolina attains perfect prediction when also adding to the three previous elements, Ag, Cd, and Cu. Further elements are necessary for bran, while Fe and Mg replace K and Ag to classify seeds. In conclusion, five elements, B, Cd, Cu, K, and Se, are the most important in distinguishing between organic and conventional agriculture; these elements also permit some differentiation among products. The method could help in fraud prevention.

Matrix-Assisted Laser Desorption Ionization Time-of-Flight Mass Spectrometry Combined with Chemometrics for Protein Profiling and Classification of Boiled and Extruded Quinoa from Conventional and Organic Crops

Quinoa is an Andean crop that stands out as a high-quality protein-rich and gluten-free food. However, its increasing popularity exposes quinoa products to the potential risk of adulteration with cheaper cereals. Consequently, there is a need for novel methodologies to accurately characterize the composition of quinoa, which is influenced not only by the variety type but also by the farming and processing conditions. In this study, we present a rapid and straightforward method based on matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS) to generate global fingerprints of quinoa proteins from white quinoa varieties, which were cultivated under conventional and organic farming and processed through boiling and extrusion. The mass spectra of the different protein extracts were processed using the MALDIquant software (version 1.19.3), detecting 49 proteins (with 31 tentatively identified). Intensity values from these proteins were then considered protein fingerprints for multivariate data analysis. Our results revealed reliable partial least squares-discriminant analysis (PLS-DA) classification models for distinguishing between farming and processing conditions, and the detected proteins that were critical for differentiation. They confirm the effectiveness of tracing the agricultural origins and technological treatments of quinoa grains through protein fingerprinting by MALDI-TOF-MS and chemometrics. This untargeted approach offers promising applications in food control and the food-processing industry.

Evaluating long-term impacts of land use/land cover changes on pollution loads at a catchment scale.

Evaluating how pollutant loads react to changes in land use/land cover (LULC) is a challenging task due to the intricate relationships among the many elements within a watershed. However, the difficulty in connecting LULC change and nonpoint source (NPS) pollution loads to streams may be lessened by combining hydrological modeling with geospatial tools and multivariate statistics. The objective of this study was to investigate the long-term effects of LULC change on NPS pollution loads in a highly human-dominated catchment, in central Ethiopia. In the study, hydrologic modeling was used to estimate the NPS parameters from multispectral Landsat images, and multivariate statistical techniques were then used to extract major LULC types that explain the variances of NPS loads between 1981 and 2020. The results demonstrated that there were human-induced LULC changes in the area, as the built-up and agricultural landscapes are rising (186.4% and 5.8%, respectively), and shrub and forest lands are decreasing (67.1% and 41%, respectively). As a result of these changes, the concentrations of nitrate (NO3), total P, total N, organic N, and organic P loads were increased by 69.41, 19.83, 18.45, 18.88, and 24.05%, respectively. Reductions in natural vegetation, as well as agriculture intensification, are the major contributors to the NPS pollutant losses to surface water sources. The result also revealed that pollution nutrients are strongly related to deforestation and agricultural land expansion. Proper adaptation strategies should be implemented to minimize the negative impact of LULC changes in the area.

LC-MS metabolomics and molecular docking approaches to identify antihyperglycemic and antioxidant compounds from Melastoma malabathricum L. Leaf

The dried leaves of Melastoma malabathricum L., locally named Karamunting or Senduduk, is traditionally consumed in many regions in Indonesia as herbal tea to cure different illnesses, including diabetes. To date, information on the compounds responsible for their antidiabetic activity is still very rare. The study aimed to identify bioactive compounds of M. malabathricum L. leaves using LC-MS based metabolomics and molecular docking approaches. The leaves brewed with different methods were subjected to LC−MS measurements and several bioactivity tests (in vivo and in vitro antihyperglycemic, and in vitro antioxidant). LC−MS data were linked to the activity data using multivariate data analysis. Molecular docking using alpha-glucosidase, alpha-amylase, and insulin receptor as protein targets was used to verify the results and study the interaction between the identified compound and protein targets. As results, isoquercetin and myricitrin were identified as compounds strongly associated with alpha-amylase inhibitors, while rutin and epicatechin were identified as alpha-glucosidase inhibitors. Quercitrin, citric acid, quercetin, epicatechin, isoquercitrin, and 7-hydroxycoumarine were strongly correlated with both antihyperglycemic and antioxidant activities. The results of metabolomics were confirmed with molecular docking studies, which showed that some of these compounds acted as competitive inhibitors, while others acted as non-competitive ones. Possible synergism between epicatechin and citric acid in their interaction with IR was detected. Metabolomics combined with molecular docking efficiently identified and confirmed several antihyperglycemic and antioxidant compounds from M. malabathricum L., leaf. This study provides scientific evidence for the traditional use of M. malabathricum L. as an antidiabetic herbal.

Extracting interpretable signatures of whole-brain dynamics through systematic comparison.

The brain's complex distributed dynamics are typically quantified using a limited set of manually selected statistical properties, leaving the possibility that alternative dynamical properties may outperform those reported for a given application. Here, we address this limitation by systematically comparing diverse, interpretable features of both intra-regional activity and inter-regional functional coupling from resting-state functional magnetic resonance imaging (rs-fMRI) data, demonstrating our method using case-control comparisons of four neuropsychiatric disorders. Our findings generally support the use of linear time-series analysis techniques for rs-fMRI case-control analyses, while also identifying new ways to quantify informative dynamical fMRI structures. While simple statistical representations of fMRI dynamics performed surprisingly well (e.g., properties within a single brain region), combining intra-regional properties with inter-regional coupling generally improved performance, underscoring the distributed, multifaceted changes to fMRI dynamics in neuropsychiatric disorders. The comprehensive, data-driven method introduced here enables systematic identification and interpretation of quantitative dynamical signatures of multivariate time-series data, with applicability beyond neuroimaging to diverse scientific problems involving complex time-varying systems.

A Study of the Association between Primary Oral Pathologies (Dental Caries and Periodontal Diseases) Using Synchrotron Molecular FTIR Spectroscopy in View of the Patient's Personalized Clinical Picture (Demographics and Anamnesis).

In this exploratory study, we searched for associations between the two most common diseases of the oral cavity-dental caries and periodontal diseases-taking into account additional factors, such as personalized clinical pictures (the individual risk factors of the patient), based on the method of a multivariate data analysis of the molecular changes in the composition of human gingival crevicular fluid (GCF). For this purpose, a set of synchrotron Fourier-transform infrared spectroscopy (FTIR) spectra of gingival crevicular fluid samples from patients with different demographics, levels of dental caries development and periodontal diseases, and the presence/absence of concomitant chronic diseases were obtained and analyzed. Using a set of techniques (v-, F-, Chi-square tests; a principal component analysis (PCA); and the hierarchical clustering of principal components (HCPCs)) implemented in the R package FactoMineR allowed us to assess the relationship between the principal components (PCs) and characteristics of the respondents. By identifying the features (vibrational modes in the FTIR spectra) that contribute most to the differentiation of the spectral dataset, and by taking into account the interrelationships between the patients' characteristics, we were able to match specific biological markers (specific molecular groups) to the two factors of interest-two types of oral pathologies. The results obtained show that the observed changes in the quantitative and qualitative composition of the modes in the infrared (IR) spectra of the GCF samples from patients with different dental caries developments and periodontal diseases present confirm the difficulty of identifying patient-specific spectral information. At the same time, different periodontal pathologies are more closely associated with other characteristics of the patients than the level of their caries development. The multivariate analysis performed on the spectral dataset indicates the need to take into account not only the co-occurrence of oral diseases, but also some other factors. The lack of this consideration (typical in lots of studies in this area) may lead to misinterpretations and consequently to a loss of data when searching for biological markers of certain oral diseases.

Gated Recurrent Unit (GRU) for Forecasting Hourly Energy Fluctuations

In the current digital era, energy use undeniably supports economic growth, increases social welfare, and encourages technological progress. Energy-related information is often presented in complex time series data, such as energy consumption data per hour or in seasonal patterns. Deep learning models are used to analyze the data. The right choice of normalization method has great potential to improve the performance of deep learning models significantly. Deep learning models generally use several normalization methods, including min-max and z-score. In this research, the deep learning model chosen is Gated Recurrent Unit (GRU) because the computational load on GRU is lighter, so it doesn't require too much memory. In addition, the GRU data is easier to train, so that it can save training time. This research phase adopts the CRISP-DM methodology in data mining as a solution commonly used in business and research. This methodology involves six stages: Business Understanding, Data Understanding, Data Preparation, Modeling, Evaluation, and Deployment. In this research, the model was obtained using five attribute selection, which applied 2 normalization methods: min-max and z-score. With this normalization, the GRU model produces the best MAPE of 3.9331%, RMSE of 0.9022, and R2 of 0.9022. However, when using z-score normalization, the model performance decreases with MAPE of 10.4332%, RMSE of 0.7602, and R2 of 0.4213. Overall, min-max normalization provides better performance in multivariate time series data analysis.

A novel standard-free detection of adulteration method for sildenafil derivatives in dietary supplements.

The rapid and accurate detection of illegal adulteration of chemical drugs into dietary supplements is a big challenge in the food chemistry field. Detection of compounds without a standard reference is even more difficult; however, this is a common situation. Here in this study, a novel "standard-free detection of adulteration" (SFDA) method was proposed and phosphodiesterase-5 inhibitor derivatives were used as an example to figure out the possibility and reliability of this SFDA method. After analysis by quadrupole coupled time of flight-tandem mass spectrometry detection and multivariable statistics, six common fragment ions were chosen to indicate whether adulteration was present or not, while 20 characteristic fragment ions indicated whether adulteration was by nitrogen-containing heterocycles or by anilines. Furthermore, the quantitative methods were conducted by high-performance liquid chromatography-tandem mass spectrometry. In a word, this strategy allows for a quick determination of dietary supplement adulteration without any need for standard materials, improving the efficacy of food safety testing.

A novel hybrid model based on multiple influencing factors and temporal convolutional network coupling ReOSELM for wind power prediction

Highly accurate wind power prediction is important for the operation and management of wind farms, energy utilization and power supply. In recent years, in order to further improve prediction accuracy, some researchers have considered utilizing factors such as wind speed and wind direction to establish prediction models. However, this often results in input redundancy due to a lack of technical means. In this paper, a multivariate prediction model for wind power based on variational mode decomposition (VMD), fuzzy entropy (FE), partial auto-correlation function and cross-correlation function (PACF&CCF, abbreviated as PC), improved artificial hummingbird algorithm (IAHA), temporal convolutional network (TCN), and regularized online sequential extreme learning machine (ReOSELM) is proposed. To address the nonlinearity and non-stationarity of the original wind power time series and decrease computational costs, the VMD combined with FE data preprocessing technique is employed. Considering the information redundancy in the multivariate data, the PC technique is utilized to conduct correlation analysis on the input data and select highly correlated features as inputs. When the input is multivariate and high-dimensional, the ReOSELM model struggles to achieve satisfactory prediction performance. Therefore, a new model based on TCN-ReOSELM is constructed. Additionally, to overcome the drawbacks of manual parameter adjustment, an IAHA algorithm is proposed, which integrates Latin hypercube sampling and chaotic local search strategies, to optimize the crucial parameters of TCN-ReOSELM. The experimental results demonstrate that the multivariate method proposed in this paper reduces the RMSE by 31% compared to the univariate method. Specifically, the proposed VMD-FE-PC-TCN-ReOSELM model reduces the RMSE, MAE, and SMAPE by 60–64%, 55–58%, and 10–36%, respectively, compared to the PC-TCN-ReOSELM model.

Concentration, Pollution Level, and Sources of Heavy Metals from Household Dust in Ramadi City, Iraq

Researchers and environmental managers have been interested in household dust pollution because exposure to heavy metals has a detrimental effect on health. This study aims to determine heavy metals concentrations (Cd, Cr, Cu, Ni, Pb and Zn), pollution level, and their potential sources in the household dust of Ramadi City, Iraq. The dust was collected from 35 houses. The atomic absorption spectrophotometry method was used to measure concentrations of heavy metals in the dust. The following was the order of the heavy metal mean concentration: Zn > Ni > Cu > Pb > Cr > Cd. These metals’ mean concentrations are below background guidelines. The descending order of the EF values for the measured metals is as follows: Ni (6.35) > Cu (5.40) > Pb (1.53 > Cr (1.24) > Zn (1.14) > Cd (0.84). Results of the calculation of CF for metals measured in house dust in Ramadi City showed values of less than 1 (CF < 1 indicating a minimum contamination level. Based on the results of multivariate statistics (correlation matrix, cluster and principle components analysis), sources of Cr, Cu, Ni and Zn are indoor anthropogenic sources while the sources of Cd and Pb are outdoor anthropogenic.

Interpretation of Hot Spots in Wuhan New Town Development and Analysis of Influencing Factors Based on Spatio-Temporal Pattern Mining

The construction of new towns is one of the main measures to evacuate urban populations and promote regional coordination and urban–rural integration in China. Mining the spatio-temporal pattern of new town hot spots based on multivariate data and analyzing the influencing factors of new town construction hot spots can provide a strategic basis for new town construction, but few researchers have extracted and analyzed the influencing factors of new town internal hot spots and their classification. In order to define the key points of Wuhan’s new town construction and promote the construction of new cities in an orderly and efficient manner, this paper first constructs a space-time cube based on the luminous remote sensing data from 2010 to 2019, extracts hot spots and emerging hot spots in Wuhan New City, selects 14 influencing factor indicators such as population density, and uses bivariate Moran’s index to analyze the influencing factors of hot spots, indicating that the number of bus stops and vegetation coverage rate are the most significant. Secondly, the disorderly multivariate logistic regression model is used to analyze the influencing factors of emerging hot spots. The results show that population density, vegetation coverage, road density, distance to water bodies, and distance to train stations are the most significant factors. Finally, based on the analysis results, some relevant suggestions for the construction of Wuhan New City are proposed, providing theoretical support for the planning and policy guidance of new cities, and offering reference for the construction of new towns in other cities, promoting the construction of high-quality cities.

Using summary tables to introduce principal component analysis in an elementary data science course

AbstractThis article presents a novel approach to introducing principal component analysis (PCA), using summary tables and descriptive statistics. Given its applicability across a variety of academic disciplines, this topic offers abundant opportunity for class discussion and activities. However, teaching PCA in an introductory class can be challenging due to the potential abstraction of multivariate datasets, and especially when students have a minimal background in statistics or data science. This method aims to help teachers bridge the gap between basic descriptive statistics and the more advanced concepts of PCA; this is done by disregarding mathematical optimization, while emphasizing the use of summary tables and the programming language R. The focus is on implementing this method in an introductory tertiary data science course; however, it may potentially be used in higher level courses, and across a variety of disciplines.

Hybrid machine learning system based on multivariate data decomposition and feature selection for improved multitemporal evapotranspiration forecasting

Evapotranspiration is an essential component of the hydrological cycle. Forecasting the reference crop evapotranspiration (ETo) using a reliable and generalized framework is crucial for agricultural operations, especially irrigation. This study was aimed at evaluating the performance of a hybrid system including the K-Best selection (KBest), multivariate variational mode decomposition (MVMD), and Machine learning (ML) models for 1-, 3-, 7-, and 10-day-ahead forecasting of the daily ETo in twelve stations of California. The analysis covered a span of 20 years, from 2003 to 2022. Three stand-alone ML models, namely Cascade Forward Neural Network (CFNN), Extreme Learning Machine (ELM), and Bagging Regression Tree (BRT) are used and were integrated with various preprocessing techniques to construct three hybrid models, i.e., MVMD-KBest-CFNN, MVMD-KBest-ELM, and MVMD-KBest-BRT. According to the results obtained in the testing phase, averaged across all stations, all three stand-alone models (CFNN, ELM, and BRT) yielded similar outcomes. In contrast, the hybrid models exhibited significantly enhanced performances compared with the standalone models, and MVMD-KBest-CFNN and MVMD-KBest-ELM models outperformed MVMD-KBest-BRT model. The BRT-based models were vulnerable to overfitting. The performance of the best models is superior compared to similar existing studies. Examining the variations across stations, it was found that the stations located further from the coast and in arid regions could be susceptible to prediction errors and necessitate more attention.

ContrAttNet: Contribution and attention approach to multivariate time-series data imputation

ABSTRACT The imputation of missing values in multivariate time-series data is a basic and popular data processing technology. Recently, some studies have exploited Recurrent Neural Networks (RNNs) and Generative Adversarial Networks (GANs) to impute/fill the missing values in multivariate time-series data. However, when faced with datasets with high missing rates, the imputation error of these methods increases dramatically. To this end, we propose a neural network model based on dynamic contribution and attention, denoted as ContrAttNet. ContrAttNet consists of three novel modules: feature attention module, iLSTM (imputation Long Short-Term Memory) module, and 1D-CNN (1-Dimensional Convolutional Neural Network) module. ContrAttNet exploits temporal information and spatial feature information to predict missing values, where iLSTM attenuates the memory of LSTM according to the characteristics of the missing values, to learn the contributions of different features. Moreover, the feature attention module introduces an attention mechanism based on contributions, to calculate supervised weights. Furthermore, under the influence of these supervised weights, 1D-CNN processes the time-series data by treating them as spatial features. Experimental results show that ContrAttNet outperforms other state-of-the-art models in the missing value imputation of multivariate time-series data, with average 6% MAPE and 9% MAE on the benchmark datasets.

Multivariate data binning and examples generation to build a Diabetic Retinopathy classifier based on temporal clinical and analytical risk factors

In this paper, we explore the possibility of exploiting retrospective clinical data from Electronic Health Records (EHR) for classification tasks in chronic patients. The different intervals, short length and high class imbalance make it unfeasible to use traditional time series techniques. The first contribution of the paper is a preprocessing method to construct a multivariate time series dataset using EHR data, which infers missing data and regularizes the data frequency. The second contribution addresses class imbalance by using domain knowledge and existing short EHR series. We synthetically extrapolate patients’ data by using similar long time series and a fuzzy-based approach. The paper addresses the problem of detection of Diabetic Retinopathy (DR). Expert domain knowledge from ophthalmologists has been used in the proposed techniques to guide the processing of time series. The novelty in that case study consists in not using eye-fundus image analysis. Instead, the proposed methods are based solely on EHR data. Several multivariate multiclass time series classifiers are used to detect the four levels of DR severity from the pre-processed data sequences. Experiments prove the quality of the sequence preprocessing techniques proposed for EHR data. Results indicate that the TapNet classifier is the best one for DR grading. Despite being tested for DR detection, the proposed data preparation methods are applicable to other diseases with similar characteristics.

Investigating swimming technical skills by a double partition clustering of multivariate functional data allowing for dimension selection

Investigating swimming technical skills by a double partition clustering of multivariate functional data allowing for dimension selection