Papers by Keyword: MLR

Abstract: This study investigates the calco-carbonic balance of drinking water in Taza, Morocco, a critical parameter for ensuring water quality and preserving distribution infrastructure. Using Principal Component Analysis (PCA) and Multiple Linear Regression (MLR) with Statistica 12, we examine the interactions between the physicochemical parameters and the Langelier Saturation Index (LSI). The study highlights the difficulty of factors influencing this balance, which is crucial for preventing scaling. Scaling can lead to reduced water flow, decreased energy efficiency, increased maintenance costs, premature equipment wear, and deterioration of water quality. The results identify the parameters impacting this balance, including temperature, total hardness, dissolved oxygen, and pH. PCA enabled us to extract valuable insights from physicochemical analyses, revealing significant correlations between these parameters and suggesting optimization strategies.The predictive model for the Langelier Saturation Index, with a determination coefficient (R² = 0.925) and a standard error (σerr = 0.07), provides a valuable tool for expecting and correcting imbalances, therefore ensuring better management of drinking water quality in Taza.

Abstract: Carcinogenicity is an important toxicological endpoint which poses a great concern being the major determinants of cancers and tumours. Anilines possess such toxic properties as they can form various electrophilic intermediates and adducts with biological systems. In the present work, the molecular descriptors of anilines have been calculated with semi-empirical AM1 and E-dragon methods, and a quantitative structure–toxicity relationships (QSTR) model for carcinogenic potency (pTD₅₀) model of anilines was developed with multiple linear regression (MLR) analysis. The validation results through the test set indicate that the proposed model is robust and satisfactory. The QSTR study suggests that the molecular structure and the electronegativity of chemicals are closely related to the Carcinogenicity.

Abstract: Nowadays, wind power has occupied the position that cannot be ignored in the electric power development. However, the accuracy of the wind speed forecast is very important for wind power generation. For proper and efficient forecast of wind speed, a novel model named WRF-SSA-MLR, which was combined by Weather Reseacher Forecast (WRF), Singular Spectrum Analysis (SSA) and Multiple Linear Regression (MLR), was proposed to predict the wind speed. The proposed model achieves 0.1318 to 7.9170 in the root mean square error (RMSE) compared to real wind speed values. For wind speed forecast, it could be a promising candidate for improving the prediction accuracy.

Abstract: Because the lance position of oxygen-enriched air and top-blown furnace is affected by lance pressure, melt temperature, melt furface,the type of slag and other factors, it is difficult accurately measured. Firstly, processing informations(excluding data,analysis correlation) to the production data.Then identified the main auxiliary variables of soft measurement model and established MLR and PCR soft prediction model of the lance position.Secondly,according to production data,through fitting and forecasting of these models, the results show that the PCR model's forecasting ability is better than the MLR model.Finally, the obtained model is embedded in WinCC software platform and further verify the feasibility of the model.The research may guide the production and lay theoretical foundations for the lance position control

Abstract: Quantitative structuretoxicity relationship (QSTR) studies play an important role in toxicity predicting, and is widely used in the study of modern compounds. Anilines represent one of the most important classes of environmental chemicals. Most of them may cause serious public health and environmental problems. The present work is to develop an effective QSTR model for mutagenicity, a toxicological endpoint which has a significant determinant of cancers, of Anilines. We calculated various descriptors and used linear regression way to select relevant parameters, and built a QSTR model that was correlation with Log P, ELUMO and heat of formation (R²=0.87, SE=0.78, Rcv²=0.867585, F=89.034). The model showed a good forecasting ability. Based on the descriptors, a further discussion was presented for the toxic mechanism. The results show that Log P value has the most important effect on anilines toxicity.

Abstract: A quantitative structure property relationship (QSPR) study was performed in this work to develop models for predicting reaction rate constants for reductive debromination of polybrominated diphenyl ethers (PBDEs) by zero-valent iron (ZVI). Both multiple linear regression (MLR) and artificial neural network (ANN) methods were employed for QSPR studies based on the experimental kinetic data of the fourteen PBDE congeners. Both the developed MLR and ANN models could give satisfactory prediction abilities, and the performance of the ANN model seems slightly better than that of the MLR model. In addition, energy of lowest unoccupied molecular orbital (ELUMO) and total energy (TE) were found to be the two relatively important variables in the ANN model via the assessment using both the Garson’s algorithm and connection weight approach.

Abstract: Quantitative structure-property relationship (QSPR) models were developed in the present work for photodegradation rate constants (kp) of fifteen individual polybrominated diphenyl ethers (PBDEs) in methanol/water (8:2) by UV light in the sunlight region. The molecular descriptors used in the QSPR models were calculated by the two semi-empirical quantum mechanical methods, RM1 and PM6, respectively. Both multiple linear regression (MLR) and artificialneural network (ANN) were applied in this study. The statistic qualities of the MLR models based on the molecular parameters obtained by RM1 and PM6 calculations were both good with the R values of 0.987 and 0.990, respectively. The QSPR model built by the ANN method with the molecular parameters calculated with PM6 is slightly better than that with RM1.

Abstract: Thermal error modeling method is an important field of thermal error compensation on NC machine tools, it is also a key for improving the machining accuracy of machine tools. The accuracy of the model directly affects the quality of thermal error compensation. On the basis of multiple linear regression (MLR) model, this paper proposes an autoregressive distributed lag (ADL) model of thermal error and establishes an accurate ADL model by stepwise regression analysis. The ADL model of thermal error is established with measured data, it proved the ADL model is available and has a high accuracy on predicting thermal error by comparing with MLR models.