Linear Regression Prediction Model Research Articles

Evaluation of keratometric and total corneal astigmatism measurements from optical biometers and anterior segment tomographers and mapping to reconstructed corneal astigmatism vector components.

To investigate different measures for corneal astigmatism in the context of reconstructed corneal astigmatism (recCP) as required to correct the pseudophakic eye, and to derive prediction models to map measured corneal astigmatism to recCP. Retrospective single centre study of 509 eyes of 509 cataract patients with monofocal (MX60P) IOL. Corneal power measured with the IOLMaster 700 keratometry (IOLMK), and Galilei G4 keratometry (GK), total corneal power (TCP2), and Alpin's integrated front (CorT) and total corneal power (CorTTP). Feedforward shallow neural network (NET) and linear regression (REG) prediction models were derived to map the measured C0 and C45 power vector components to the respective recCP components. Both the NET and REG models showed superior performance compared to a constant model correcting the centroid error. The mean squared prediction errors for the NET/REG models were: 0.21/0.33 dpt for IOLMK, 0.23/0.36 dpt for GK, 0.24/0.35 for TCP2, 0.23/0.39 dpt for CorT and 0.22/0.36 dpt for CorTTP respectively (training data) and 0.27/0.37 dpt for IOLMK, 0.26/0.37 dpt for GK, 0.38/0.42 dpt for TCP2, 0.35/0.36 dpt for CorT, and 0.44/0.45 dpt for CorTTP respectively on the test data. Crossvalidation with model optimisation on the training (and validation) data and performance check on the test data showed a slight overfitting especially with the NET models. Measurement modalities for corneal astigmatism do not yield consistent results. On training data the NET models performed systematically better, but on the test data REG showed similar performance to NET with the advantage of easier implementation.

Predictive Modeling of Global Temperature Change Based on Artificial Intelligence Algorithms

As the greenhouse effect intensifies, global temperature changes are increasingly affecting all aspects of economic life. The purpose of this paper is to analyze the influencing factors of temperature to establish an analytical model of global temperature change. The Berkeley Earth Climate Database provides access to the data needed for the study. Pearson correlation analysis was used to quantify the strength of the correlations between temperature, location and other factors. After that, global temperatures were modelled and predicted by artificial intelligence algorithms such as ARIMA and linear regression prediction models, respectively. The results test that the model has good stability and can achieve accurate prediction function. The model can better help predict and respond to global temperature changes. The application of this AI algorithm can be used as a reference in other fields.

Based on the improved fuzzy analytic hierarchy and the TSE-MLR model energy consumption prediction of university: A case study

The accurate prediction of building energy consumption on university campuses is a significant research area. Current studies often focus on predicting the energy consumption of specific building areas or individual equipment, and typically consider only one factor, limiting the accuracy and applicability of the predictions. This study introduces the Time Segmented Energy-Multiple Linear Regression (TSE-MLR) prediction model, which integrates the improved fuzzy analytic hierarchy and the multiple linear regression algorithm. The model is compared with traditional (MLR, BP) and advanced (RNN) models, and their various indexes are discussed and analyzed. By collecting meteorological and energy consumption data from the study site over the past 12 years, the key factors affecting energy consumption on the university campus were identified using the improved fuzzy analytic hierarchy. Subsequently, the TSE-MLR model was trained using energy consumption data from 2010 to 2016 and validated using data from 2017 to 2019. The prediction results of the TSE-MLR model were compared with those obtained through Multiple linear regression, BP neural networks, and RNN. The results demonstrated that the TSE-MLR model significantly reduced the prediction error by 13.8 % and exhibited higher accuracy compared to the other models. Therefore, the TSE-MLR model introduced in this study offers a new and effective approach to predicting university energy consumption and supporting energy management using existing data from university building operations across different periods.

Soft tissue movements after mandibular reconstruction using the vascularized iliac flap: patterns and predictions.

To investigate soft-to-hard tissue response following mandibular reconstruction and to develop a predictive model for projecting soft tissue movement. In this retrospective study, 18 patients receiving mandibular reconstruction using a vascularized iliac flap were enrolled. Various indicators for characterizing the movement of tissues were considered to identify the effective predictors for projecting soft tissue movements. Face-region-specific linear regression models for prediction were constructed and evaluated. The arithmetic mean of hard tissue movement in an extended area had the strongest correlation with the movement of the focal soft tissue, while the arithmetic mean in a regional area (Ram) was a more effective predictor. The linear regression model using Ram, global extrema and distances between them as the predictors performed the best in the lower margin of the face, with an average error of 1.51 ± 1.38mm. Soft tissue movement in the alveolar process was not correlated with the existence of dentition, only can be predicted by the soft tissue movement below it. The area of the masseter was strongly correlation with Ram, but no other factors. An accurate prediction of soft tissue movements in the lower margin and the alveolar process of the face can be achieved by considering hard tissue and adjacent soft tissue movements. No effective predictor in the masseter area was identified. We investigated the relationship between hard tissue movements and the soft tissue responses in the facial area. Through building predictive models for projecting postoperative soft tissue movements, we derive insights for the aesthetic outcome of face surgeries. This study was registered on the Chinese Clinical Trial Registry (registration number: ChiCTR2100054103).

Transport characteristics of heavy metals in the soil-atmosphere-wheat system in farming areas and development of multiple linear regression predictive model

Heavy metal accumulation in agricultural products has become a major concern. Previous studies have focused on the transport of heavy metals from the soil and their accumulation in crops. However, recent studies revealed that wheat leaves, ears, and awns can also transport and accumulate heavy metals. Wheat grains can be influenced by two sources of heavy metals: soil contamination and atmospheric deposition. To comprehend the transport characteristics of heavy metals in soil, atmospheric deposition, and wheat, 37 samples each for wheat rhizosphere soil, wheat roots, stems, leaves, and grains were collected. Fifteen samples of atmospheric dry deposition and atmospheric wet deposition were collected from Linshu County (northern area), China. Based on the test data, the characteristics of heavy metals and their distribution in the study area were analyzed. Migration patterns of heavy metals in crops from different sources were investigated using Pearson correlation and redundancy analysis. Finally, a predictive model for heavy metals in wheat grains was developed using multiple linear regression analysis. Significant disparities in the distribution of heavy metals existed among wheat roots, stems, leaves, and grains. The coefficient of variation of heavy metals in atmospheric deposition was relatively high, indicating discernible spatial patterns influenced by human activities. Notably, a positive correlation was observed between the concentration of heavy metals in wheat grains and atmospheric deposition of Hg, Cd, and Pb. Conversely, Zn and Ni levels in wheat grains were significantly negatively associated with soil Zn, Ni, pH, and OM content. The contribution of heavy metal elements from different sources varied in their impact on the grain's heavy metal content. Specifically, atmospheric deposition was the primary source of Hg and Pb in wheat grains, while Cd, Ni, Cu, and Zn were predominantly derived from soil. Using a multiple linear regression model, we could accurately predict Hg, Pb, Cd, Ni, Zn, and As concentrations in crop grains. This model can facilitate quantitative evaluation of ecological risk of heavy metals accumulation in crops in the study area.

Improving prediction of linear regression models by integrating external information from heterogeneous populations: James-Stein estimators.

We consider the setting where (1) an internal study builds a linear regression model for prediction based on individual-level data, (2) some external studies have fitted similar linear regression models that use only subsets of the covariates and provide coefficient estimates for the reduced models without individual-level data, and (3) there is heterogeneity across these study populations. The goal is to integrate the external model summary information into fitting the internal model to improve prediction accuracy. We adapt the James-Stein shrinkage method to propose estimators that are no worse and are oftentimes better in the prediction mean squared error after information integration, regardless of the degree of study population heterogeneity. We conduct comprehensive simulation studies to investigate the numerical performance of the proposed estimators. We also apply the method to enhance a prediction model for patella bone lead level in terms of blood lead level and other covariates by integrating summary information from published literature.

Recent Advancements and Trends of Topical Drug Delivery Systems in Psoriasis: A Review and Bibliometric Analysis.

Psoriasis is an immune-mediated inflammatory skin disease where topical therapy is crucial. While various dosage forms have enhanced the efficacy of current treatments, their limited permeability and lack of targeted delivery to the dermis and epidermis remain challenges. We reviewed the evolution of topical therapies for psoriasis and conducted a bibliometric analysis from 1993 to 2023 using a predictive linear regression model. This included a comprehensive statistical and visual evaluation of each model's validity, literature profiles, citation patterns, and collaborations, assessing R variance and mean squared error (MSE). Furthermore, we detailed the structural features and penetration pathways of emerging drug delivery systems for topical treatment, such as lipid-based, polymer-based, metallic nanocarriers, and nanocrystals, highlighting their advantages. This systematic overview indicates that future research should focus on developing novel drug delivery systems characterized by enhanced stability, biocompatibility, and drug-carrying capacity.

Non-invasive biomonitoring of polar bear feces can be used to estimate concentrations of metals of concern in traditional food.

The Arctic faces increasing exposure to environmental chemicals such as metals, posing health risks to humans and wildlife. Biomonitoring of polar bears (Ursus maritimus) can be used to quantify chemicals in the environment and in traditional foods consumed by the Inuit. However, typically, these samples are collected through invasive or terminal methods. The biomonitoring of feces could be a useful alternative to the current metal monitoring method within the Arctic. Here, we aim to 1) quantify the relationship between concentrations of metals in the feces and tissues (muscle, liver, and fat) of polar bears using predictive modeling, 2) develop an easy-to-use conversion tool for use in community-based monitoring programs to non-invasively estimate contaminant concentrations in polar bears tissues and 3) demonstrate the application of these models by examining potential exposure risk for humans from consumption of polar bear muscle. Fecal, muscle, liver, and fat samples were harvested from 49 polar bears through a community-based monitoring program. The samples were analyzed for 32 metals. Exploratory analysis indicated that mean metal concentrations generally did not vary by age or sex, and many of the metals measured in feces were positively correlated with the internal tissue concentration. We developed predictive linear regression models between internal (muscle, liver, fat) and external (feces) metal concentrations and further explored the mercury and methylmercury relationships for utility risk screening. Using the cross-validated regression coefficients, we developed a conversion tool that contributes to the One Health approach by understanding the interrelated health of humans, wildlife, and the environment in the Arctic. The findings support using feces as a biomonitoring tool for assessing contaminants in polar bears. Further research is needed to validate the developed models for other regions in the Arctic and assess the impact of environmental weathering on fecal metal concentrations.

Research on Automatic Pricing and Replenishment Decision of Vegetable Products Based on Machine Learning

In fresh food supermarkets, due to the short shelf life of vegetable products and the deterioration of products over time, unsold products on the same day cannot be sold the next day. Therefore, supermarkets need to make automatic pricing and replenishment decisions on a daily basis based on historical sales and demand. All product data for this study were provided by the 2023 Higher Education Society Cup National College Student Mathematical Modeling Competition.This study aims to explore the relationship between commodity sales and wholesale prices and pricing using five machine learning methods: neural networks, linear regression, decision tree regression, random forest regression, and XGBOOST regression, in order to predict sales and wholesale prices for the next week. By comparing the performance of five machine learning methods, it was found that the XGBOOST regression prediction model is the most suitable for predicting the daily sales of goods. The linear regression prediction model is the most accurate for predicting the wholesale prices of "chili", "flower and leaf", "edible fungi", "cauliflower", and "eggplant" goods, while the random forest regression prediction model is the most accurate for predicting the wholesale prices of "aquatic rhizome" goods. By constructing a dynamic programming model and using genetic algorithms to solve, automatic pricing and replenishment decisions for various goods within the next week can be obtained.

Prediction of Rock's Brittleness and Dynamic Properties Utilizing Effective Artificial Intelligence Approaches

This research aims to determine the brittleness index (BI) and engineering properties of limestone specimens. In addition, this study evaluates the effect of moisture on the developed models to predict the BI and shear wave velocity (Vs), based on the point load index (Is50), dry and saturated tensile strength (Ts-d and Ts-s), and porosity. Gaussian process regression (GPR), multilayer feed-forward neural network (MFFNN), and multiple linear regression (MLR) predictive models were utilized. Microscopic examination of the limestone specimens revealed that calcite is the predominant mineral. It was observed that samples with higher calcite content exhibited greater brittleness and strength properties while having lower porosity. The results obtained from the MLR analysis demonstrated that it is possible to accurately forecast the brittleness index (BI), as well as the dry and saturated shear wave velocities (Vs-d and Vs-s) at the specific sites under investigation. The moisture effect on developed models showed that Vs prediction in dry conditions (Vs-d) was less accurate compared to the saturated conditions (Vs-s). Conversely, the relationships developed for estimating the BI in dry conditions exhibited higher accuracy. The analysis of all model assumptions using MLR indicated that the models could be reliably utilized. However, the MFFNN and GPR methods were found to be more conservative in estimating these properties. Moreover, the study identified the best transfer function and training algorithm for predicting Vs and BI. The evaluation metrics, such as R2 and RMSE revealed that GPR demonstrated higher precision compared to MFFNN and MLR.

Regional Carbon Emission Forecasting Model Considering Temporal Characteristics and Analysis of Peak Carbon Assessment in Typical Scenarios

Leading high-quality economic development with green development is an important development concept for Chinese-style modernization, and achieving the dual-carbon goals of carbon peaking and carbon neutrality is a key initiative to promote high-quality economic and social development. In order to explore the correlation between carbon emissions and population, economy and energy consumption, and to propose a reasonable carbon emissions prediction model, this paper proposes an evaluation index system integrating fossil energy consumption and new energy consumption by analyzing the sample data obtained from the research in a hierarchical, industry- and sector-specific manner. And then, we compare the reliability of the results obtained from the multiple linear regression prediction model based on the above index system and the Kaya carbon emissions prediction model, respectively. By comparing the reliability of the results obtained from the carbon emission multiple linear regression prediction model based on the above index system and the Kaya carbon emission prediction model, it is initially determined that the Kaya model will be used for regional carbon emission prediction. In order to optimize the sensitivity of the model to the temporal Characterization of the indicators, a carbon emission prediction model that incorporates the temporal characteristics of the data is set up. Lastly, the carbon emission is predicted according to the three typical scenarios analyzed according to the intensity of the macro-policy, and the year of carbon peak and the year of carbon peak are analyzed for the scenarios. Three typical scenarios are analyzed according to the intensity of macro policies, and carbon emissions are predicted under different scenarios, and the year in which each scenario reaches the carbon peak is analyzed, as well as the correlation with the implementation of related policies and energy use, with a view to providing support for the planning of the dual-carbon target value and the path of achieving it, as well as the formulation of related policies.

Early Bloomer and Future Master: Getting to the Top of the Modern Badminton World

The road to the top is never easy. This study investigated whether the career trajectories of top-level men’s and women’s badminton players could be predicted by their age at first major debut and the time taken to reach the top 150. Data from the BWF from October 2009 to October 2022 were analyzed using a predictive linear regression model with Bayesian inference, adjusting for anthropometric factors. The results suggested that age at debut influences career high rankings among over 120 elite players. Additionally, observations challenged the Matthew effect in early-career success for young players, proposing that inherent talent may be more significant than early nurturing. The study also examined the potential impact of relative age and early specialization in sports.

Taguchi method and neural network for efficientβ‐ketoenaminesynthesis in deionized water

AbstractThe optimization of performance parameters, in particular the yield of the synthesis reaction of β‐enaminones in demineralized water, is crucial to improve their efficiency and accuracy. In this report, we investigate the optimization of the synthesis of β‐ketoenamines in deionized water by controlling several parameters such as reaction time, temperature, amine equivalent, acid percentage, and stirring rate. An orthogonal L16(45) network was created using Taguchi's approach, allowing for the best possible parameters. To forecast the contribution of each parameter, analysis of variance (ANOVA) techniques are also used. Multiple linear and nonlinear regression (MLR, MNLR) and multilayer perception artificial neural network (MLP‐ANN) predictive models were developed. Analysis of the results led to optimized design parameters, with time = 6 h, temperature = 25°C, amine equivalent = 1.5, acid percentage = 20%, and stirring rate = 1000 rpm, leading to a maximum yield of 63%. ANOVA analysis revealed that temperature, stirring rate, acid percentage, and time are the parameters with the greatest influence. The least sensitive parameter is the amine equivalent. The two main interactions are temperature * acid % and amine equivalent * rpm. The MLP‐ANN predictions are in good agreement with the experimental values, resulting in a higherR2compared to the quadratic regression model and the MLR model. By using molecular docking studies, the produced compounds' biological activity was investigated. Some of the synthesized compounds appear to be interesting and could be used for therapeutic applications. The results of this study give us insight into the gentle, cost‐effective, and biologically active synthesis of β‐enaminones in deionized water.

Artificial neural network and multiple linear regression modeling for predicting thermal transmittance of plain-woven cotton fabric

The present research compares a machine learning model with a statistical model, with specific emphasis on artificial neural networks and multiple linear regression models. The aim of this study is to forecast the thermal transmittance of a plain-woven cotton fabric using input data such as thread density measured in ends per inch, picks per inch, and fabric thickness. The artificial neural network is built using a network with feed-forward backpropagation, and the MATLAB software’s training function trainlm is used to modify its weight and basic values based on Levenberg–Marquardt optimization techniques. The sigmoid transfer function is used to set the layer output and measure network performance in terms of the root mean squared error, mean absolute error percentage, and coefficient of determination which were determined. For the artificial neural network prediction model, the root mean squared error and mean absolute error percentage were 1.05 and 3.132%, respectively, while the coefficient of determination was 0.9307. In contrast, the multiple linear regression prediction model had root mean squared error and mean absolute error percentage values of 2.98 and 8.97%, respectively, along with a coefficient of determination of 0.4727. The results reveal that the artificial neural network model outperforms the multiple linear regression model, showing superior accuracy and robustness in capturing the intricate interactions between important fabric parameters (ends per inch, picks per inch, and thickness) and thermal transmittance values. This research emphasizes the efficiency of artificial neural network modeling as a superior tool for forecasting thermal transmittance in textile applications rather than employing the time-consuming trial-and-error process for delivering significant insights for material engineering and energy-efficient design.

Research on pricing and replenishment strategy of superstore goods based on linear regression and gray prediction models

The pricing and replenishment strategy of various commodities in superstores have a significant impact on the revenue. This paper aims to address the automatic decision-making problem for pricing and replenishment of vegetable commodities in superstores. It applies the linear regression model to predict commodity prices, aiming to maximize the revenue of superstores. Additionally, it utilizes the gray prediction model to forecast replenishment volume. Based on marketing data from 2020 to 2023 for vegetable goods at a specific superstore, this study provides the replenishment quantity and pricing strategy for each vegetable category from July 1-7, 2023, as well as for each individual vegetable item on July 1, 2023.

Predictive modelling of total operating room time for Laparoscopic Cholecystectomy using pre-operatively known indicators to guide accurate surgical scheduling in a critical access hospital

The financial margin of rural and critical access hospitals highly depends on their surgical volume. An efficient operating room is necessary to maximise profit and minimise financial loss. OR utilisation is a crucial OR efficiency metric requiring accurate case duration estimates. The patient's age, ASA, BMI, Mallampati score, previous surgery, the planned surgery, the surgeon, the assistant's level of experience and the severity of the patient's disease are also associated with operative duration. Although complex machine-learning models are accurate in operative prediction, they are not always available in resource-limited hospitals. Laparoscopic cholecystectomy (LC) is one of the most common surgical procedures performed and is one of the few procedures performed at critical access and rural hospitals. The accurate estimation of the operative duration of LC is essential for efficient OR utilisation. We hypothesise that a multivariate linear regression prediction model can be constructed from a set of pre-operatively known, easily collected variables to maximise OR utilisation and improve operative scheduling accuracy for LC. We further hypothesise that this model can be implemented in resource-limited environments, such as critical access hospitals.

Optimization with Adaptive Learning: A Better Approach for Reducing SSE to Fit Accurate Linear Regression Model for Prediction

Optimization with Adaptive Learning: A Better Approach for Reducing SSE to Fit Accurate Linear Regression Model for Prediction

Modeling and Detection Scheme for Zero-Dynamics Attack on Wind Power System

In a wind farm, wind power system is key to maintaining continuous generation of point energy. Since Zero-Dynamics attack poses a major risk to future wind farms, investigation on the effect of Zero-Dynamics attack against wind power systems has both theoretical and practical importance. In this paper, we consider a scenario in which a wind power system is subject to Zero-Dynamics attack in power generation chain. Firstly, we transform a semi-direct D-PMSG into an equivalent circuit to establish mathematical model of wind power system under Zero-Dynamics attack. Then, we analyze and prove the advantages and limitations of Zero-Dynamics attack by divide the equation of state into a stable part and an unstable part with the help of system relative degree ρ. In addition, to better achieve desired attack effect, adjustments are made for Zero-Dynamics attack based on its limitations. Finally, we combine an Multiple Linear Regression (MRL) predictive control model with Byrnes-Isidori normal form to protect wind power system. The theoretical results were verified on designed small wind power system.

Research on energy-saving strategies in the architectural design of the Eastern Railway in the era of artificial intelligence

Abstract This paper takes the air conditioning energy consumption in railroad building design as the main object of analysis and uses an orthogonal experimental design method to study the influencing factors and effects of energy consumption in the Middle East railroad construction design. Stochastic working condition simulation is carried out for the typical models of passenger stations in five climate zones to obtain the data set of the passenger station energy consumption model, and the multiple linear regression prediction models of air conditioning energy consumption in different regions are established. Combined with the hour-by-hour average outdoor temperature, hour-by-hour passenger flow and lighting and related equipment data in the waiting hall of a city in July, the air-conditioning cold load of railroad passenger stations is simulated and analyzed. Explore the implementation effect of energy-saving strategies for railroad passenger station air-conditioning system, and for the traditional PID control strategy, put forward the control strategy based on model prediction, analyze the change of centralized air-conditioning system cold load and the change of the actual performance coefficient of the chiller unit under different control strategies. The data show that compared with the traditional PID control, the model prediction control reduces the total energy consumption by 0.2063, which realizes the energy-saving optimization of the air conditioning system for railroad building design.

Semi-parametric single-index predictive regression models with cointegrated regressors

This paper considers the estimation of a semi-parametric single-index regression model that allows for nonlinear predictive relationships. This model is useful for predicting financial asset returns, whose observed behaviour resembles a stationary process, if the multiple nonstationary predictors are cointegrated. The presence of cointegrated regressors imposes a single-index structure in the model, and this structure not only balances the nonstationarity properties of the multiple predictors with the stationarity properties of asset returns but also avoids the curse of dimensionality associated with nonparametric regression function estimation. An orthogonal series expansion is used to approximate the unknown link function for the single-index component. We consider the constrained nonlinear least squares estimator of the single-index (or the cointegrating) parameters and the plug-in estimator of the link function, and derive their asymptotic properties. In an empirical application, we find some evidence of in-sample nonlinear predictability of U.S. stock returns using cointegrated predictors. We also find that the single-index model in general produces better out-of-sample forecasts than both the historical average benchmark and the linear predictive regression model.