Generic Prediction Research Articles

Application of Machine Learning for Generic Mill Liner Wear Prediction in Semi-Autogenous Grinding (SAG) Mills

This study explores the application of machine learning techniques for predicting generic mill liner wear in semi-autogenous grinding (SAG) mills used in mineral processing. Various models were developed and compared using data from 143 liner measurements across 36 liner cycles from ten different SAG mills. The research initially focused on individual mill modeling, employing simple linear regression, first-order kinetic approach, Multiple Linear Regression (MLR), tree-based methods (Decision Trees, Random Forests, XGBoost), and Multilayer Perceptron (MLP). Results showed that simple linear regression provided sufficient accuracy, with other methods only slightly improving performance. This study then developed a combined model using data from multiple mills. MLR and advanced machine learning techniques were applied for this generic model, with XGBoost emerging as the most successful. In the interpolation scenario involving a mill similar to those in the training data, the XGBoost model achieved a mean absolute percentage error (MAPE) of 5.27%. For the extrapolation scenario, with a mill larger than those in the training set, the MAPE increased slightly to 6.12%. These results demonstrate the potential of machine learning approaches in creating effective generic models for mill liner wear prediction. However, this study also highlights the potential for improving predictive models by incorporating additional key parameters such as liner and ball material properties.

LIMO-GCN: a linear model-integrated graph convolutional network for predicting Alzheimer disease genes.

Alzheimer's disease (AD) is a complex disease with its genetic etiology not fully understood. Gene network-based methods have been proven promising in predicting AD genes. However, existing approaches are limited in their ability to model the nonlinear relationship between networks and disease genes, because (i) any data can be theoretically decomposed into the sum of a linear part and a nonlinear part, (ii) the linear part can be best modeled by a linear model since a nonlinear model is biased and can be easily overfit, and (iii) existing methods do not separate the linear part from the nonlinear part when building the disease gene prediction model. To address the limitation, we propose linear model-integrated graph convolutional network (LIMO-GCN), a generic disease gene prediction method that models the data linearity and nonlinearity by integrating a linear model with GCN. The reason to use GCN is that it is by design naturally suitable to dealing with network data, and the reason to integrate a linear model is that the linearity in the data can be best modeled by a linear model. The weighted sum of the prediction of the two components is used as the final prediction of LIMO-GCN. Then, we apply LIMO-GCN to the prediction of AD genes. LIMO-GCN outperforms the state-of-the-art approaches including GCN, network-wide association studies, and random walk. Furthermore, we show that the top-ranked genes are significantly associated with AD based on molecular evidence from heterogeneous genomic data. Our results indicate that LIMO-GCN provides a novel method for prioritizing AD genes.

Predicting metro incident duration using structured data and unstructured text logs

Predicting metro incident duration is crucial for passengers and transit operators to choose appropriate response strategies. Most existing research focuses on structured data, the rich information embedded within unstructured incident logs is often neglected. This paper incorporates a probabilistic topic model tailored for short texts, the biterm topic model, into the generic incident duration prediction models. By capturing text co-occurrence patterns through Bayesian inference, the biterm topic model extracts hidden topics from incident narratives, and each topic serves as a condensed summary of detailed incident causes and countermeasures. These extracted topics are then combined with structured information to serve as predictors. We validated our model using five years of incident data from the Hong Kong Mass Transit Railway across two scenarios: with all incident information available and information revealed over time . The results demonstrate that our method significantly improves the prediction accuracy, particularly for incidents lasting longer than 30-minute.

Analysis and prediction of the Horizon 2020 R&D&I collaboration network

This study presents a novel link prediction model for the analysis and prediction of collaboration networks in the context of H2020 projects. This model incorporates various databases and expands the list of variables used in the analysis. In this study, the accuracy and interpretability of both generic and nongeneric link prediction algorithms are examined through a comprehensive analysis. Additionally, collaborative groups that exhibit stronger linkages than initially anticipated are examined, and their prevalence across different regions is evaluated. In this study, it was found that the collaboration network within H2020 has significant concentration and fragmentation levels. Moreover, several key factors that strongly influence collaboration, including prior and concurrent project involvement, earned contribution, and the economic and technological characteristics of participating businesses, were identified. This study additionally demonstrated that cooperative communities exhibit geographical concentration in western Europe, whereas nongeneric models offer improved predictive accuracy for identifying relationships. This research contributes to the field of cooperation network analysis and offers guidance to decision-makers on strategies to promote future collaborations in the context of FP.

Requiem to “proof of inflation” or sourced fluctuations in a non-singular bounce

Popular wisdom suggests that measuring the tensor to scalar ratio r on CMB scales is a “proof of inflation” since one generic prediction is a scale-invariant tensor spectrum while alternatives predict r that is many orders of magnitude below the sensitivity of future experiments. A bouncing Universe with sourced fluctuations allows for nearly scale-invariant spectra of both scalar and tensor perturbations challenging this point of view. Past works have analyzed the model until the bounce, under the assumption that the bounce will not change the final predictions. In this work, we discard this assumption. We explicitly follow the evolution of the Universe and fluctuations across the bounce until reheating. The evolution is stable, and the existence of the sourced fluctuations does not destroy the bounce. The bounce enhances the scalar spectrum while leaving the tensor spectrum unchanged. The enhancement depends on the duration of the bounce — a shorter bounce implies a larger enhancement. The model matches current observations and predicts any viable tensor-to-scalar ratio r ≲ 10-2, which may be observed in upcoming CMB experiments. Hence, a measurement of r will no longer be a “proof of inflation”, and a Sourced Bounce is a viable paradigm with distinct predictions.

Prediction method of centrifugal pump running in turbine mode based on losses analysis

Interest of Pump as Turbine (PAT) is growing with diverse applications in engineering. Usually, the data of PATs are not available in the hands of pump manufacturers. Therefore, performance prediction methods appear as an important research area of PATs. The current prediction methods reposed on expensive, inaccurate and time consuming experimental methods. In the scope of this work, a generic and robust prediction method is built up for a centrifugal impeller PAT. The most significant hydraulic losses were derived in PAT mode, these are namely, the shock losses at the impeller inlet, the swirling losses at the impeller outlet and the impeller wall frictional losses. The Euler head, the available total head and the hydraulic efficiency were computed as well. The global efficiency was computed taking into account the machine mechanical and volumetric efficiencies, enabling therefore to perform comparison of the new prediction method with experimental, computational fluid dynamic (CFD), Rossi and Perez performances prediction methods. From where it resulted a good agreement between the given prediction methods for the entire range of operation, confirming the robustness and the applicability of the developed prediction method. The relative difference between the new prediction method and CFD data and between the new prediction method and experimental data remained higher for lower discharge conditions, notably for extreme part load conditions, where a small error could result in very high relative difference.

ProkDBP: Toward more precise identification of prokaryotic DNA binding proteins.

Prokaryotic DNA binding proteins (DBPs) play pivotal roles in governing gene regulation, DNA replication, and various cellular functions. Accurate computational models for predicting prokaryotic DBPs hold immense promise in accelerating the discovery of novel proteins, fostering a deeper understanding of prokaryotic biology, and facilitating the development of therapeutics targeting for potential disease interventions. However, existing generic prediction models often exhibit lower accuracy in predicting prokaryotic DBPs. To address this gap, we introduce ProkDBP, a novel machine learning-driven computational model for prediction of prokaryotic DBPs. For prediction, a total of nine shallow learning algorithms and five deep learning models were utilized, with the shallow learning models demonstrating higher performance metrics compared to their deep learning counterparts. The light gradient boosting machine (LGBM), coupled with evolutionarily significant features selected via random forest variable importance measure (RF-VIM) yielded the highest five-fold cross-validation accuracy. The model achieved the highest auROC (0.9534) and auPRC (0.9575) among the 14 machine learning models evaluated. Additionally, ProkDBP demonstrated substantial performance with an independent dataset, exhibiting higher values of auROC (0.9332) and auPRC (0.9371). Notably, when benchmarked against several cutting-edge existing models, ProkDBP showcased superior predictive accuracy. Furthermore, to promote accessibility and usability, ProkDBP (https://iasri-sg.icar.gov.in/prokdbp/) is available as an online prediction tool, enabling free access to interested users. This tool stands as a significant contribution, enhancing the repertoire of resources for accurate and efficient prediction of prokaryotic DBPs.

The dynamics of vacuum, gravity and matter: Implications on the fundamental constants

The possibility that the vacuum energy density (VED) [Formula: see text] could be time dependent in the expanding Universe is intuitively more reasonable than just a rigid cosmological constant for the entire cosmic history. The dynamics of [Formula: see text] as a function of the Hubble rate, [Formula: see text], most likely contributes to alleviate cosmological problems and tensions, having also implications on the so-called fundamental ‘constants’ of Nature, which should be slowly drifting with the cosmic expansion owing to the fluctuations of the quantum vacuum. This includes the gravitational ‘constant’ G, but also the gauge and Yukawa couplings as well as the particle masses themselves (both of dark matter and baryonic matter). The subtle exchange of energy involved is the basis for the “micro and macro connection”. Herein, I discuss not only this connection as a possibility but also show that it is in fact a generic prediction of QFT in cosmological spacetime which is fully compatible with general covariance. This fact has not been pointed out until recently when an appropriate renormalization framework for the VED has been found which is free from the usual conundrums associated with the cosmological constant problem.

Electromagnetic Signatures of Mirror Stars

Mirror stars are a generic prediction of dissipative dark matter (DM) models, including minimal atomic DM and twin baryons in the mirror twin Higgs model. Mirror stars can capture regular matter from the interstellar medium through extremely suppressed kinetic mixing interactions between the regular and the dark photon. This accumulated “nugget” will draw heat from the mirror star core and emit highly characteristic X-ray and optical signals. In this work, we devise a general parameterization of mirror star nugget properties that is independent of the unknown details of mirror star stellar physics, and use the Cloudy spectral synthesis code to obtain realistic and comprehensive predictions for the thermal emissions from optically thin mirror star nuggets. We find that mirror star nuggets populate an extremely well-defined and narrow region of the Hertzsprung–Russell diagram that only partially overlaps with the white dwarf population. Our detailed spectral predictions, which we make publicly available, allow us to demonstrate that optically thin nuggets can be clearly distinguished from white dwarf stars by their continuum spectrum shape, and from planetary nebulae and other optically thin standard sources by their highly exotic emission-line ratios. Our work will enable realistic mirror star telescope searches, which may reveal the detailed nature of DM.

On the Actionability of Outcome Prediction

Predicting future outcomes is a prevalent application of machine learning in social impact domains. Examples range from predicting student success in education to predicting disease risk in healthcare. Practitioners recognize that the ultimate goal is not just to predict but to act effectively. Increasing evidence suggests that relying on outcome predictions for downstream interventions may not have desired results. In most domains there exists a multitude of possible interventions for each individual, making the challenge of taking effective action more acute. Even when causal mechanisms connecting the individual's latent states to outcomes are well understood, in any given instance (a specific student or patient), practitioners still need to infer---from budgeted measurements of latent states---which of many possible interventions will be most effective for this individual. With this in mind, we ask: when are accurate predictors of outcomes helpful for identifying the most suitable intervention? Through a simple model encompassing actions, latent states, and measurements, we demonstrate that pure outcome prediction rarely results in the most effective policy for taking actions, even when combined with other measurements. We find that except in cases where there is a single decisive action for improving the outcome, outcome prediction never maximizes "action value", the utility of taking actions. Making measurements of actionable latent states, where specific actions lead to desired outcomes, may considerably enhance the action value compared to outcome prediction, and the degree of improvement depends on action costs and the outcome model. This analysis emphasizes the need to go beyond generic outcome prediction in interventional settings by incorporating knowledge of plausible actions and latent states.

Generic and accurate prediction of retention times in liquid chromatography by post–projection calibration

Retention time predictions from molecule structures in liquid chromatography (LC) are increasingly used in MS–based targeted and untargeted analyses, providing supplementary evidence for molecule annotation and reducing experimental measurements. Nevertheless, different LC setups (e.g., differences in gradient, column, and/or mobile phase) give rise to many prediction models that can only accurately predict retention times for a specific chromatographic method (CM). Here, a generic and accurate method is present to predict retention times across different CMs, by introducing the concept of post–projection calibration. This concept builds on the direct projections of retention times between different CMs and uses 35 external calibrants to eliminate the impact of LC setups on projection accuracy. Results showed that post–projection calibration consistently achieved a median projection error below 3.2% of the elution time. The ranking results of putative candidates reached similar levels among different CMs. This work opens up broad possibilities for coordinating retention times between different laboratories and developing extensive retention databases.

Systematic development and validation of a predictive model for major postoperative complications in the Peri-operative Quality Improvement Project (PQIP) dataset.

Complications are common following major surgery and are associated with increased use of healthcare resources, disability and mortality. Continued reliance on mortality estimates risks harming patients and health systems, but existing tools for predicting complications are unwieldy and inaccurate. We aimed to systematically construct an accurate pre-operative model for predicting major postoperative complications; compare its performance against existing tools; and identify sources of inaccuracy in predictive models more generally. Complete patient records from the UK Peri-operative Quality Improvement Programme dataset were analysed. Major complications were defined as Clavien-Dindo grade ≥ 2 for novel models. In a 75% train:25% test split cohort, we developed a pipeline of increasingly complex models, prioritising pre-operative predictors using Least Absolute Shrinkage and Selection Operators (LASSO). We defined the best model in the training cohort by the lowest Akaike's information criterion, balancing accuracy and simplicity. Of the 24,983 included cases, 6389 (25.6%) patients developed major complications. Potentially modifiable risk factors (pain, reduced mobility and smoking) were retained. The best-performing model was highly complex, specifying individual hospital complication rates and 11 patient covariates. This novel model showed substantially superior performance over generic and specific prediction models and scores. We have developed a novel complications model with good internal accuracy, re-prioritised predictor variables and identified hospital-level variation as an important, but overlooked, source of inaccuracy in existing tools. The complexity of the best-performing model does, however, highlight the need for a step-change in clinical risk prediction to automate the delivery of informative risk estimates in clinical systems.

Generic prediction model of moisture content for maize kernels by combing spectral and color data through hyperspectral imaging

Moisture content (MC) is an important index to measure the quality of maize kernels. This study aims to construct a generic prediction model and optimize the characteristic variables for efficiently predicting the MC of maize kernels across various varieties. Visible/near-infrared hyperspectral imaging (HSI) within the wavelength range of 374.98 - 1038.79 nm is employed. The 270 samples containing 18 varieties at five periods over two years are collected. Spectral and color feature data based on the H, S, and V color channels of maize are extracted. Partial least squares regression (PLSR) and principal component regression (PCR) are used to establish MC prediction models. Backward interval least squares regression (Bi-PLS), uninformative variables elimination (UVE), and successive projections algorithm (SPA) are jointly (Bi-PLS, UVE, Bi-PLS-SPA, UVE-SPA, and SPA) used to select characteristic wavelengths. The MC prediction models are developed using whole wavelengths, characteristic wavelengths, color features, and fused data. The results suggest that the optimal PLSR model is based on fused data of color features and characteristic wavelengths selected by UVE-SPA from S-G smoothing spectral data (HSV+S-G-UVE-SPA-PLSR). The Rc and Rp are 0.9804 and 0.9835, respectively. The RMCEc and RMCEp are 1.6889% and 1.6523%, respectively. The RPD is 5.22. The optimal prediction model can quickly measure MC for different maize kernels, guiding diverse application scenarios to enhance the quality of maize kernels and processing efficiency.

Enhancing Generic Reaction Yield Prediction through Reaction Condition-Based Contrastive Learning.

Deep learning (DL)-driven efficient synthesis planning may profoundly transform the paradigm for designing novel pharmaceuticals and materials. However, the progress of many DL-assisted synthesis planning (DASP) algorithms has suffered from the lack of reliable automated pathway evaluation tools. As a critical metric for evaluating chemical reactions, accurate prediction of reaction yields helps improve the practicality of DASP algorithms in the real-world scenarios. Currently, accurately predicting yields of interesting reactions still faces numerous challenges, mainly including the absence of high-quality generic reaction yield datasets and robust generic yield predictors. To compensate for the limitations of high-throughput yield datasets, we curated a generic reaction yield dataset containing 12 reaction categories and rich reaction condition information. Subsequently, by utilizing 2 pretraining tasks based on chemical reaction masked language modeling and contrastive learning, we proposed a powerful bidirectional encoder representations from transformers (BERT)-based reaction yield predictor named Egret. It achieved comparable or even superior performance to the best previous models on 4 benchmark datasets and established state-of-the-art performance on the newly curated dataset. We found that reaction-condition-based contrastive learning enhances the model's sensitivity to reaction conditions, and Egret is capable of capturing subtle differences between reactions involving identical reactants and products but different reaction conditions. Furthermore, we proposed a new scoring function that incorporated Egret into the evaluation of multistep synthesis routes. Test results showed that yield-incorporated scoring facilitated the prioritization of literature-supported high-yield reaction pathways for target molecules. In addition, through meta-learning strategy, we further improved the reliability of the model's prediction for reaction types with limited data and lower data quality. Our results suggest that Egret holds the potential to become an essential component of the next-generation DASP tools.

Transient Tip Clearance Prediction Model Considering Transient Radial Temperature Distribution of Discs in a Gas Turbine Engine

Abstract The tip clearance of turbomachinery components, such as compressor and turbine, significantly impacts the gas turbine engine performance, including the thrust and fuel consumption. Therefore, an appropriate prediction of the turbomachinery tip clearance is indispensable in improving the accuracy of engine performance evaluations. The traditional 1D transient tip clearance prediction model does not consider the radial temperature distribution of the disk (also called disc). However, as the engine performance keeps improving, the radial temperature gradient of the disk also increases. In order to maintain the prediction accuracy within an acceptable range, a generic transient tip clearance prediction model is developed in this paper. Based on the special treatment of the axial and radial convective heat transfer boundary regions of the disk, a 2D transient heat transfer model is established, and the radial temperature of the disk is predicted. At the same time, the interaction effect of the thermal expansion between the two different layers of the shroud is studied to improve the prediction accuracy further. In addition, the centrifugal deformation of the blade and disk with variable-thickness characteristics is considered in the new transient model. Compared with the traditional 1D transient prediction model, the model developed in this paper further improves the predicting accuracy of the transient tip clearance, which is verified by the high-fidelity fluid–solid–thermal coupling method. The prediction model provides a useful tool for evaluating the transient tip clearance. It may also benefit the future design strategy of active clearance control.

Wood volume is overestimated in the Brazilian Amazon: Why not use generic volume prediction methods in tropical forest management?

The Amazon has a range of species with high potential for sustainable timber harvesting, but for them to be utilized globally, the merchantable wood volume must be accurately quantified. However, since the 1950s, inadequate methods for estimating merchantable timber volumes have been employed in the Amazon, and Brazilian Government agencies still require some of them. The natural variability of the Amazon Forest provides an abundance of species of different sizes and shapes, conferring several peculiarities, which makes it necessary to use up-to-date and precise methods for timber quantification in Amazon Forest management. Given the employment of insufficient estimation methods for wood volume, this study scrutinizes the disparities between the actual harvested merchantable wood volume and the volume estimated by the forest inventory during the harvesting phase across five distinct public forest areas operating under sustainable forest management concessions. We used mixed-effect models to evaluate the relationships between inventory and harvested volume for genera and forest regions. We performed an equivalence test to assess the similarity between the volumes obtained during the pre-and post-harvest phases. We calculated root mean square error and percentage bias for merchantable volume as accuracy metrics. There was a strong tendency for the 100% forest inventory to overestimate merchantable wood volume, regardless of genus and managed area. There was a significant discrepancy between the volumes inventoried and harvested in different regions intended for sustainable forest management, in which only 22% of the groups evaluated were equivalent. The methods currently practiced by forest companies for determining pre-harvest merchantable volume are inaccurate enough to support sustainable forest management in the Amazon. They may even facilitate the region's illegal timber extraction and organized crime.

Multivariate optimization of hyperspectral imaging for adulteration detection of ground beef: Towards the development of generic algorithms to predict adulterated ground beef and for digital sorting

Due to its nutritional value, and taste, minced beef meat (MBM) is highly valuable therefore is susceptible to economically motivated adulteration. High-throughput adulteration detection methods are required to guarantee the authenticity and safety of the MBM. Hyperspectral imaging (HSI) in the 400–1000 nm spectral range, in tandem with multivariate analysis, was evaluated to demonstrate the feasibility of developing generic models to detect minced beef meat adulterated with other types of meats. Partial least squares regression (PLSR) method, the ensemble Monte Carlo variable selection method (EMCVS), a range of spectral pre-treatments, and combinations of any two of them were evaluated to predict the amount of MBM in each prepared sample scanned. The beef prediction model was developed using data from MBM adulterated with minced chicken and turkey meats and validated with data from MBM adulterated with pork meat at adulteration levels ranging from 0 to 51% at approximately 3% increments. Good prediction results were obtained using the EMCVS, on the asymmetric least squares (AsLs) + Standard Normal variate (SNV) pre-treated reflectance spectra (Rp2 0.96, RMSEP 2.9% and RPD 5.4) with 23 selected wavelengths. An independent set of MBM samples adulterated with lamb and duck meat were then prepared similarly at concentrations ranging from 3 to 21% to test the developed method. Good results were obtained in the full range (R2p 0.94, RMSEp 3.92%, RPDp 3.83) and excellent prediction results (R2p 0.95, RMSEp 2.78%, RPDp 5.29) were obtained with 9 optimum wavelengths (484, 508, 593, 596, 602, 736, 739, 924 and 960 nm). PLS discriminant analysis (PLS-DA), the EMCVS and spectral pre-treatments were used to discriminate MBM from other adulterants and achieved almost perfect classification for calibration, cross-validation, and prediction using 12 selected spectral bands. This study developed a high-throughput generic minced meat adulteration prediction and discrimination method based on selected spectral bands that can be used to further develop low-cost portable sensors for the digital sorting of adulterated MBM, and demonstrates the feasibility to develop generic models to detect minced beef meat adulterated with other types of meat.

A test of gravity with Pulsar Timing Arrays

A successful measurement of the Stochastic Gravitational Wave Background (SGWB) in Pulsar Timing Arrays (PTAs) would open up a new window through which to test the predictions of General Relativity (GR). We consider how these measurements might reveal deviations from GR by studying the overlap reduction function — the quantity that in GR is approximated by the Hellings-Downs curve — in some sample modifications of gravity, focusing on the generic prediction of a modified dispersion relation for gravitational waves. We find a distinct signature of such modifications to GR — a shift in the minimum angle of the angular distribution — and demonstrate that this shift is quantitatively sensitive to any change in the phase velocity. In a given modification of gravity, this result can be used, in some regions of parameter space, to distinguish the effect of a modified dispersion relation from that due to the presence of extra polarization modes.

Gauged SU(3)F and loop induced quark and lepton masses

We investigate a local SU(3)F flavour symmetry for its viability in generating the masses for the quarks and charged leptons of the first two families through radiative corrections. Only the third-generation fermions get tree-level masses due to specific choice of the field content and their gauge charges. Unprotected by symmetry, the remaining fermions acquire non-vanishing masses through the quantum corrections induced by the gauge bosons of broken SU(3)F. We show that inter-generational hierarchy between the masses of the first two families arises if the flavour symmetry is broken with an intermediate SU(2) leading to a specific ordering in the masses of the gauge bosons. Based on this scheme, we construct an explicit and predictive model and show its viability in reproducing the realistic charged fermion masses and quark mixing parameters in terms of not-so-hierarchical fundamental couplings. The model leads to the strange quark mass, ms ≈ 16 MeV at MZ, which is ~2.4σ away from its current central value. Large flavour violations are a generic prediction of the scheme which pushes the masses of the new gauge bosons to 103 TeV or higher.

Prediction model for multiblock data with or without intermediate blocks. Estimation by PLS Path Modeling

An important issue in modern science remains the integration of multiple data sources, considered as datasets divided into blocks of variables measured on the same set of observations. When analyzing these blocks, the path modeling approach aims to take into account a specific pattern of directed relations between them. This structure, usually set up from prior knowledge, leads to a path diagram in which the block’s components are linked to one another. Until path modeling has gained popularity in recent decades, the issue of prediction has been surprisingly neglected in the literature. To fill this gap, we propose a prediction framework dedicated to the path modeling approach. Indeed, prediction for path modeling methods is of paramount interest for: (i) selecting the optimal model with cross-validation (e.g., optimal model specification and dimension), (ii) comparing the model with other relevant ones, (iii) or predicting new observations. Here, we propose an explicit formulation of a prediction model in the context of a path modeling structure. This generic prediction model is estimated using the PLS Path Modeling algorithm, denoted PMM-PLSPM (=Prediction Model for Multiblock data estimated by PLS Path Modeling). Nevertheless, its versatility makes it possible to adapt the approach to any other composite-based methods. Two estimation strategies (PLSPM-like and PLSR-like) are proposed for prediction, both being viewed as the structural model at the variable level. At the same time, the components of multiple blocks – obtained by a relevant deflation strategy – are handled to take advantage of the full multidimensional potential of blocks. The proposed prediction strategies are applied and evaluated on real data sets.