Pre-selection Method Research Articles

A new portfolio approach integrating three-way decision and Encoder–Decoder network

In the stock market, it can be challenging to predict stock trends and identify stocks with investment potential due to uncertainty and risk. Accurate predictive models and reliable asset preselection methods are essential for portfolio management. This paper proposes an integration of the deep learning model, the three-way decision (3WD) theory, and the Mean-Variance (MV) method for asset preselection and optimal asset allocation. In the initial stage, an attention-based Encoder–Decoder model is used to predict asset returns and calculate volatility. It outperforms other benchmark models regarding relative error, absolute error, and directional accuracy. The subsequent stage involves selecting stocks based on expected returns and volatility characteristics. The 3WD theory is then used to identify the delayed decision set for implementing additional investment strategies to hedge portfolio risk and enhance returns. Finally, the MV model is employed for investment portfolio optimization. The proposed model is validated using a substantial dataset from the A-share market’s SSE 50 Index from January 2010 to December 2022. Results indicate that the proposed model outperforms benchmark models in terms of daily average returns, annualized Information ratio, Sortino ratio, and other aspects under certain trading costs. However, further exploration is required to establish a mechanism that limits the quantity of preselected assets to ensure its broad applicability.

Parameter Study for Establishing a Synchronizer Control Strategy in Tractor Dual-Clutch Transmission

Research on dual-clutch transmissions in agricultural tractors is on the rise. Dual-clutch transmissions provide a smooth driving experience and easy operation compared to manual transmissions, as they eliminate power interruptions. Accordingly, there is ongoing research to improve the conventional synchronizer pre-selection method. The study introduces a new control strategy to enhance shifting quality by focusing on shift force and timing. Specifically, the goal is to optimize the impact of pre-selection-induced shock and delayed shift time, minimizing discomfort for the driver. According to the experimental results, it was determined that factors influencing the impact of preselect engagement on gear shift shock were absent, while factors capable of affecting shift time were identified in the form of shift force. However, it was concluded that optimizing shift force and shift timing based on the number of shifts can enhance the quality of gear shifting. Even during preselect disengagement, the factor influencing gear shift shock was identified as the shift timing, with no discernible factor affecting shift time. Nevertheless, it was judged that optimizing shift force and shift timing based on the number of shifts can lead to an improvement in gear shifting quality. In conclusion, the control strategy determined through the parameter study is expected to improve the shifting quality of tractors equipped with dual-clutch transmission. These findings have the potential to support efficient and convenient tractor operation in the agricultural sector.

Research on power system flexibility enhancement method based on pre-selection of structure adjustment strategy

The flexibility enhancement method of power system usually focuses on node flexible resource, without considering the transmission process. It may result in some flexible resources not being transferred, thus affecting the accuracy and effectiveness of the flexibility enhancement. In this paper, a flexibility enhancement optimization model based on pre-selection method of structure adjustment strategy is proposed, which can improve system flexibility with guaranteed availability of flexible resources. Firstly, the flexibility enhancement pre-selection method based on power system topology adjustment mathematical formulation is proposed to select the lines suitable for topology adjustment. Secondly, a flexibility enhancement optimization model considering economy and available flexibility is developed, which takes the maximum available flexibility index as the objective function and ensures the system flexible and economical by cost-floating constraint. The available flexibility deviation constraint corrects the system flexibility supply to guarantee the accuracy of available flexibility. Finally, the effectiveness of the proposed method is investigated through the case study on modified IEEE 24-bus and IEEE 118-bus system.

Topology optimization of multi-material structures considering anisotropic yield strengths

Multi-material structures offer the superior performance and broader designable freedom, and have gained increasing interest benefitting from the advances in additive manufacturing (AM) technique. However, the effect of anisotropic yield strength of these AM-fabricated parts on topology optimization of multi-material structures has so far received little attention. In the current work, an anisotropic strength-based topology optimization method for multi-material structures is developed for maximizing strength performance under total mass constraint, which allows for the consideration of directional dependence and tension-compression asymmetry of yield strengths based on the Tsai-Wu yield criterion. Another challenge for multi-material structures design is that the possible permutations of numerous candidate materials are unlimited, it is impossible that all combinations of multiple candidate materials are tested for the optimal strength performance of multi-material structures. To solve this problem, we propose a preselection method for the inclusion or exclusion of candidate materials based on their elastic properties, yield strengths and densities before optimization. Afterwards, our topology optimization method can automatically find the optimal combination of volume fractions of different candidate materials. Several 2D and 3D numerical examples are investigated for minimizing the maximum of failure indexes of strength for fixed total mass. The advantages of multi-material structures are verified by comparing to single-material designs. The effectiveness of the preselection method of candidate materials is also verified.

Accurate genomic selection using low-density SNP panels preselected by maximum likelihood estimation

Genomic selection (GS) poses a challenge for the prediction of the genomic estimated breeding value (GEBV) using a low-density SNP panel. Several methods have been proposed for SNP preselection. However, these methods often suffer from either significant computational complexity or erratic accuracy in GS. In this study, we developed an approach called MLE-rank based on maximum likelihood estimation to preselect a set of SNPs for GS. First, we generated 90 simulated datasets and compared the performance of MLE-rank with uniform distribution and preselection based on a genome-wide association study (GWAS). For simulated datasets, compared to uniform distribution, both MLE-rank and GWAS preselection reduced the SNP density by a factor of 10 while maintaining prediction accuracy. Additionally, compared to the other two methods, MLE-rank's prediction accuracy was significantly improved with the medium- and high-heritability datasets. Then, we further evaluated these three preselection approaches using real disease-resistant phenotypes of leopard coral grouper (Plectropomus leopardus) and Japanese flounder (Paralichthys olivaceus). We found that the 3 k SNPs preselected by MLE-rank had a stable and effective prediction effect. The uniform distribution requires 70 k, while the GWAS preselection method requires 3 k (P. leopardus) and 50 k (P. olivaceus) to achieve similar prediction accuracy. Finally, we evaluated the prediction accuracy of MLE-rank using candidate populations of flounders and their progeny survival rates, with uniform distribution and GWAS preselection as benchmarks. In the results for this dataset, MLE-rank was found to have the same predictive effect for low-density SNP panels as it did for high-density SNPs, regardless of whether GWAS preselection or uniform distribution was used. Taken together, the results we have observed indicate that we have ensured that MLE-rank does not reduce prediction accuracy for any of the datasets. MLE-rank showed superior performance in reducing the number of SNPs. Moreover, we observed a relative standard deviation in prediction accuracy when using a low density of SNPs selected by MLE-rank compared to a high density determined through a uniform distribution strategy. In conclusion, MLE-rank not only reduces the number of SNPs used for GS but also exhibits high predictive accuracy. This could potentially lead to a decrease in genotyping costs and promote the wider application of GS in fish breeding.

High-precision and efficiency diagnosis for polymer electrolyte membrane fuel cell based on physical mechanism and deep learning

As a nonlinear and dynamic system, the polymer electrolyte membrane fuel cell (PEMFC) system requires a comprehensive failure prediction and health management system to ensure its safety and reliability. In this study, a data-driven PEMFC health diagnosis framework is proposed, coupling the fault embedding model, sensor pre-selection method and deep learning diagnosis model. Firstly, a physical-based mechanism fault embedding model of PEMFC is developed to collect the data on various health states. This model can be utilized to determine the effects of different faults on cell performance and assist in the pre-selection of sensors. Then, considering the effect of fault pattern on decline, a sensor pre-selection method based on the analytical model is proposed to filter the insensitive variable from the sensor set. The diagnosis accuracy and computational time could be improved 3.7% and 40% with the help of pre-selection approach, respectively. Finally, the data collected by the optimal sensor set is utilized to develop the fault diagnosis model based on 1D-convolutional neural network (CNN). The results show that the proposed health diagnosis framework has better diagnosis performance compared with other popular diagnosis models and is conducive to online diagnosis, with 99.2% accuracy, higher computational efficiency, faster convergence speed and smaller training error. It is demonstrated that faster convergence speed and smaller training error are reflected in the proposed health diagnosis framework, which can significantly reduce computational costs.

Evaluating natural background levels of heavy metals in shallow groundwater of the Pearl River Delta via removal of contaminated groundwaters: Comparison of three preselection related methods

Assessing natural background levels (NBLs) in groundwater is a global concern. Knowledge on groundwater NBLs in urbanized areas is challenging due to the impact of complex human activities. Preselection related methods are common ones for assessing groundwater NBLs. The present study used three preselection related methods to assess groundwater heavy metals (lead, zinc, barium) NBLs in four groundwater units of the Pearl River Delta (PRD) where urbanization continues, and to identify the best one for assessing groundwater NBLs in urbanized areas. Here, methods include a preselection method (method–P), a preselection dominated method (method–PD), and a statistic dominated method (method–SD). Results showed that the method–PD was better than other two methods for assessing groundwater NBLs of heavy metals in the PRD. This is supported by the evidence that differences among heavy metals concentrations in various land-use types in residual datasets formed by the method–PD were insignificant. NBLs of lead in groundwater units I to IV assessed by the method–PD were 2.8 μg/L, 5.9 μg/L, 5.8 μg/L, and 2.6 μg/L, respectively. NBLs of zinc in groundwater units I to IV assessed by the method–PD were 30 μg/L, 180 μg/L, 160 μg/L, and 100 μg/L, respectively. NBLs of barium in groundwater units I to IV assessed by the method–PD were 120 μg/L, 120 μg/L, 90 μg/L, and 50 μg/L, respectively. Compared to the method–PD, the method–SD often underestimates groundwater NBLs of heavy metals because of using the experiential evaluation for residual datasets. The method–P also has an inaccurate evaluation of groundwater NBLs of heavy metals in comparison with the method–PD, owing to both of using the experiential evaluation and the absence of a function for outliers test. The method–P combining with an outliers test would be better than itself for assessing groundwater NBLs. Therefore, the method–PD is the first choice to be recommended for assessing groundwater NBLs in urbanized areas such the PRD. However, this method should not be taken into account for assessing groundwater NBLs in areas where groundwater Cl/Br mass ratios are invalid. Instead, the method–SD and the method–P combining with one outliers test may be choices, because no constraint for these two methods.

Superhydrophobic and superoleophilic melamine sponges impregnated with deep eutectic solvents for oil spill cleanup

The extensive extraction of oil from the bottom of seas and oceans and its transportation by tankers increase the risk of potential environmental disasters associated with hydrocarbon fractions entering water reservoirs. Therefore, this paper presents the preparation of a simple impregnation of a melamine sponge with deep eutectic solvents (DES), which can be obtained from natural sources, including coconut oil, palm kernel oil, seed oils, mammalian milk, and substances commonly used in pharmacological and physiological studies. A preselection method based on the COSMO-RS model was used to select the most suitable DES among 760 solvents. The selected DES were prepared and impregnated on melamine sponge (MS). Structural and surface analyses were performed by DES-MS. The process parameters that affect the sorption capacity of DES-MS and the affinity tests were performed for specific oil groups. The highest absorption capacities were obtained for MS impregnated with tetraethylammonium chloride and decanoic acid in a 1:2 M ratio. The prepared new DES-MS were characterized by very high sorption capacities ranging from 40.9 to 99.8 g/g, as well as superhydrophobicity (water contact angle 151°), superoleophobicity (oil contact angle 0°), low density (0.177 g/cm3), high porosity (88.25 %), high thermal, mechanical, and chemical stability, and the possibility of multiple regenerations without loss of sorption capacity. In addition, the preparation of DES-MS is simple and involves little time and consumption of toxic organic solvents. These results indicate that DES-MS may be a “green” and low-cost alternative to current approaches for removing oil from water.

Use Test of Automated Machine Learning in Cancer Diagnostics.

Our aim is to investigate the added value of automated machine learning (AutoML) for potential future applications in cancer diagnostics. Using two important diagnostic questions, the non-invasive determination of IDH mutation status and ATRX status, we analyze whether it is possible to use AutoML to develop models that are comparable in performance to conventional machine learning models (ML) developed by experts. For this purpose, we develop AutoML models using different feature preselection methods and compare the results with previously developed conventional ML models. The cohort used for our study comprises T2-weighted MRI images of 124 patients with histologically confirmed gliomas. Using AutoML, we were able to develop sophisticated models in a very short time with only a few lines of computer code. In predicting IDH mutation status, we obtained a mean AUC of 0.7400 and a mean AUPRC of 0.8582. ATRX mutation status was predicted with very similar discriminatory power, with a mean AUC of 0.7810 and a mean AUPRC of 0.8511. In both cases, AutoML was even able to achieve a discriminatory power slightly above that of the respective conventionally developed models in a very short computing time, thus making such methods accessible to non-experts in the near future.

A 325F 2 Physical Unclonable Function Based on Contact Failure Probability With Bit Error Rate < 0.43 ppm After Preselection With 0.0177% Discard Ratio

Preselection methods in which unstable bitcells are discarded are widely used to enhance the stability of the physical unclonable function (PUF). If the discard range of the preselection is too narrow, all unstable bitcells cannot be discarded, whereas, if the discard range is too wide, the PUF circuit area increases because the number of redundant bitcells should increase. Moreover, the discard range of most previous PUFs is not determined by theoretical analysis but rather by repeated measurements, which requires a long test time and cannot determine the accurate discard range to guarantee the PUF stability. This article demonstrates an efficient preselection method to determine the discard range through a theoretical analysis of the PUF entropy source for the PUF based on contact failure probability (contact PUF). After this preselection with the discard ratio of 0.0177%, no PUF responses changed among 2 367 040 bits from 1138 chips, i.e., bit error rate <inline-formula> <tex-math notation="LaTeX">$&lt;$</tex-math> </inline-formula>0.43 ppm, in seven harsh environmental tests regulated by the Joint Electron Device Engineering Council (JEDEC) standard. The proposed preselection method also enables the contact PUF to be implemented in a small area due to the low discard ratio. A small-area design methodology of contact PUF bitcells is also proposed. The PUF bitcell area per bit is 325<inline-formula> <tex-math notation="LaTeX">$F^{2}$</tex-math> </inline-formula>, the smallest area compared to previous PUFs. The contact PUF with the proposed preselection applied is fabricated in 100-nm CMOS technology. It achieves 0.4999 of uniqueness and 0.9942 of Shannon entropy and passes all applicable randomness tests stipulated by the National Institute of Standards and Technology (NIST) SP800-22.

Haplotype-tagged SNPs improve genomic prediction accuracy for Fusarium head blight resistance and yield-related traits in wheat

Key messageLinkage disequilibrium (LD)-based haplotyping with subsequent SNP tagging improved the genomic prediction accuracy up to 0.07 and 0.092 for Fusarium head blight resistance and spike width, respectively, across six different models.Genomic prediction is a powerful tool to enhance genetic gain in plant breeding. However, the method is accompanied by various complications leading to low prediction accuracy. One of the major challenges arises from the complex dimensionality of marker data. To overcome this issue, we applied two pre-selection methods for SNP markers viz. LD-based haplotype-tagging and GWAS-based trait-linked marker identification. Six different models were tested with preselected SNPs to predict the genomic estimated breeding values (GEBVs) of four traits measured in 419 winter wheat genotypes. Ten different sets of haplotype-tagged SNPs were selected by adjusting the level of LD thresholds. In addition, various sets of trait-linked SNPs were identified with different scenarios from the training-test combined and only from the training populations. The BRR and RR-BLUP models developed from haplotype-tagged SNPs had a higher prediction accuracy for FHB and SPW by 0.07 and 0.092, respectively, compared to the corresponding models developed without marker pre-selection. The highest prediction accuracy for SPW and FHB was achieved with tagged SNPs pruned at weak LD thresholds (r2 < 0.5), while stringent LD was required for spike length (SPL) and flag leaf area (FLA). Trait-linked SNPs identified only from training populations failed to improve the prediction accuracy of the four studied traits. Pre-selection of SNPs via LD-based haplotype-tagging could play a vital role in optimizing genomic selection and reducing genotyping costs. Furthermore, the method could pave the way for developing low-cost genotyping methods through customized genotyping platforms targeting key SNP markers tagged to essential haplotype blocks.

Hydrogeochemical characterization and natural background level determination of selected inorganic substances in groundwater from a semi-confined aquifer in Midwestern Burkina Faso, West Africa.

Hydrogeochemical processes that govern selected inorganic substances distribution in a semi-confined aquifer were characterized using traditional hydrogeochemical approaches and natural background levels (NBLs). Saturation indices and bivariate plots were used to investigate the effects of water-rock interactions on natural evolution of the groundwater chemistry, whereas Q-mode hierarchical cluster analysis and one-way analysis of variance classified the groundwater samples into three distinct groups. To highlight the groundwater status, NBLs and threshold values (TVs) of the substances were calculated using pre-selection method. Piper's diagram showed that the Ca-Mg-HCO3 water type was the only hydrochemical facies of the groundwaters. Although all samples, except a borewell with a high NO3- concentration, had major ion and transition metal concentrations within the World Health Organization's recommended guideline values for drinking water, Cl-, NO3- and PO43- exhibited scattered distribution patterns, reflecting their nonpoint anthropogenic sources in the groundwater system. The bivariate and saturation indices revealed that silicate weathering and possible gypsum and anhydrite dissolution contributed to the groundwater chemistry. In contrast, NH4+, FeT and Mn abundance appeared to be influenced by redox conditions. Strong positive spatial correlations between pH, FeT, Mn and Zn suggested that mobility of these metals was controlled by pH. The relative high F- concentrations in lowland areas may imply the impact of evaporation on this ion's abundance. Contrary to TVs of HCO3-, those of Cl-, NO3-, SO42-, F- and NH4+ were below the guideline values, confirming the influence of chemical weathering on the groundwater chemistry. Based on the present findings, further studies that take into account more inorganic substances are required for NBLs and TVs determination in the area, thereby setting up a robust sustainable management plan for the regional groundwater resources.

Numerical matching of supercharger for gasoline engine based on three-line working condition method

The traditional study of matching gasoline engine and supercharger mainly focuses on the external characteristic conditions, but the reasonable matching of external characteristic conditions cannot guarantee the reasonable matching of supercharger and gasoline engine in the whole operating range. Therefore, to address the limitation of studying only the external characteristic conditions in the supercharger matching process, a thermodynamic model of a direct injection supercharged gasoline engine was established, based on which the matching between the supercharger and the gasoline engine in the whole operating range was studied by the three-line condition method. The results show that the matching performance of the supercharger in the whole engine operating range can be significantly improved by using the three-line working condition method to match the supercharger. The three-line method takes into account both full-load and part-load engine conditions, and can visually evaluate the operating efficiency of the supercharger over the entire engine operating range, as well as predict the likelihood of surging, blocking and overspeeding of the supercharger over the entire engine operating range, overcoming the limitation of studying only the external characteristic conditions and providing a numerical evaluation method for pre-selection and optimization of the supercharger.

Average variance portfolio optimization using machine learning-based stock price prediction case of renewable energy investments

With the progress of time series prediction, several recent developments in machine learning have shown that the integration of prediction methods into portfolio selection is a great opportunity to structure investment decisions in the renewable energy industry. In this paper, we propose a novel approach to portfolio formation strategy based on a hybrid machine learning model that combines a convolutional neural network (CNN) and long-term bidirectional memory (BiLSTM) with robust input characteristics obtained from Huber’s location for stock prediction and the mean-variance (MV) Markowitz model for optimal portfolio construction. Specifically, this study first applies a prediction method for stock pre-selection to ensure high-quality stock inflows for portfolio formation. Then, the predicted results are integrated into the MV model. To comprehensively demonstrate the superiority of the proposed model, we used two portfolio models, the MV model and the equal-weighted (1/N) portfolio model, with LSTM, BiLSTM and CNN-BiLSTM, and used them as references. Between January 2016 and December 2021, historical data from the Stock Exchange of Thailand 50 Index (SET50) was collected for the study. Experience shows that integrating stock pre-selection can improve VM performance, and the results of the proposed method show that they outperform comparison models in terms of Sharpe ratio, average return and risk.

A Unidirectional Voltage Vector Preselection Strategy for Optimizing Model Predictive Torque Control With Discrete Space Vector Modulation of IPMSM

Discrete space vector modulation (DSVM) is employed to synthesize large number of voltage vectors (VVs) to improve the steady-state performance of predictive torque control (PTC). However, enumerating all the VVs in the prediction process increases the computation load of the motor drive. To address this issue, this article proposes a unidirectional VV preselection (UVVP) method for optimizing the performance of DSVM-based PTC of ac motors. The proposed UVVP method divides both voltage space vector diagram (VSVD) and flux space trajectory into twelve 30° based sectors to reduce the impact of candidate VVs with negative effects. At each flux position, the UVVP strategy can restrict the nearest candidate VVs in the 30° based VSVD region within the circular flux trajectory. This is achieved by using the speed direction to avoid the candidate VVs, which are in the same flux sector for reverse flux rotation, hence resulting in significant flux and torque ripple reduction. The main benefit of the proposed method is its simplicity since it only requires the flux sector and speed information as well as it can generate the VV enumerations online while ensuring the reduction of computation burden.

Factors Controlling Natural Background Levels of Ammonium and Iodide in Shallow Groundwater of Coastal Aquifers, South China

Assessing natural background levels (NBLs) in groundwater is crucial for evaluating groundwater pollution and the use of groundwater resources in coastal areas. This study assessed NBLs of iodide and ammonium in the shallow groundwater of the Pearl River Delta (PRD) by using a preselection method with Grubbs’ test, and discussed factors controlling NBLs in various groundwater units. Here, the preselection method consists of Cl/Br mass ratios versus Cl concentrations and the oxidation capacity, and the PRD is divided into four groundwater units. Results showed that NBL-iodide in groundwater unit A was 0.14 mg/L and &gt;2 times greater than that in other groundwater units. Similarly, NBL-ammonium in groundwater unit A was 0.32 mg/L and also &gt;2 times greater than that in other groundwater units. The release of iodide from both of organic-iodine in the vadose zone and iodine-rich minerals in aquifer sediments were the two main sources for the higher NBL-iodide in groundwater unit A compared to other units. By contrast, the occurrence of ammonium from organic-nitrogen in the vadose zone was the major source for the higher NBL-ammonium in groundwater unit A compared with the other units. Soluble iodide resulted from the mineralization of organic-iodine in Quaternary marine formation, and the release of iodide accompanied with reductive dissolution of iodide-loaded Fe (oxyhydr) oxides in aquifer sediments was the main driving force controlling the higher NBL-iodide in groundwater unit A compared with the other units. By contrast, the release of soluble ammonium from the mineralization of organic-nitrogen in marine formation entering into groundwater was the main driving force controlling the higher NBL-ammonium in groundwater unit A relative to the other units. These results enhance the knowledge on groundwater NBLs in coastal areas and improve groundwater resources management in coastal areas such as the PRD.

Assessing natural background levels of geogenic contaminants in groundwater of an urbanized delta through removal of groundwaters impacted by anthropogenic inputs: New insights into driving factors

Knowledge on driving forces controlling natural background levels (NBLs) of geogenic contaminants (GCs) in groundwater of coastal urbanized areas are still limited because of complex hydrogeological conditions and anthropogenic activities. This study assesses NBLs of two GCs including arsenic (As) and manganese (Mn) in four groundwater units of the Pearl River Delta (PRD) with large scale urbanization by using a preselection method composed of the chloride/bromide mass ratio versus chloride concentration and the oxidation capacity with the combination of Grubbs' test. More importantly, driving factors controlling NBLs of As/Mn in groundwater of the PRD are discussed. Results showed that groundwater As/Mn concentrations in residual datasets were independent of land-use types, while those in original datasets in different land-use types were distinct because of various human activities, indicating that the used preselection method in this study is valid for NBLs-As/Mn assessment in groundwater of the PRD. NBL-As in coastal-alluvial aquifers was >6 times that in other groundwater units. NBL-Mn in coastal-alluvial aquifers was 1.4 times that in alluvial-proluvial aquifers, and both were >4 times that in other two groundwater units. High NBLs-As/Mn in coastal-alluvial aquifers is mainly attributed to reduction of FeMn oxyhydr(oxides) induced by mineralization of organic matter in Quaternary sediments. Elevated pH also contributes higher NBL-As in coastal-alluvial aquifers. By contrast, higher NBL-Mn in alluvial-proluvial aquifers than in other two groundwater units mainly ascribes to reduction of FeMn oxyhydr(oxides) in Quaternary sediments triggered by irrigation of reducing river waters. In addition, more occurrence of As/Mn-rich sediments and the infiltration of As/Mn-rich river water are also important factors for high NBLs-As/Mn in coastal-alluvial aquifers. This study shows that revealing natural driving factors of GCs-rich groundwater in coastal urbanized areas on the basis of identification of contaminated groundwaters via the used preselection methods is acceptable.

A double pre-selection method for natural background levels assessment in coastal groundwater bodies

To evaluate the chemical status of groundwater bodies (GWB) according to the European Groundwater Directive, EU Member States are required to take into account natural background levels (NBLs) where needed. Assessing the NBLs in coastal GWBs is complicated by seawater intrusion which can be amplified by groundwater withdrawals increasing the salinization of such groundwater systems. This paper proposes a new method for the NBLs assessment in coastal areas based on a double pre-selection (PS) with fixed/dynamic limits. A case study in the Apulia region, located in southeastern Italy, is proposed, where we investigated four adjacent GWBs which form the complex karst, fractured Murgia aquifer, hosted in the Jurassic-Cretaceous carbonate platform, bounded by two seas and sustained by saltwater of marine intrusion in the coastal areas. Data related to 139 monitoring stations (MSs) of the regional groundwater monitoring network were used. The first PS, “static”, based on a fixed limit of anthropogenic contamination markers (NO3 and NH4), allows for the elimination of MSs impacted by human activities. On these, the second PS, “dynamic”, based on the identification of Cl anomalous values, allows for the identification of additional MSs affected by saline contamination. The residual dataset of MSs was used for the definition of NBLs of Cl, SO4, F and B. A statistical comparison with historical Cl observations finally allowed us to verify if the salinity of current groundwater is representative of pristine conditions. The calculated NBLs of salinity parameters are higher for the two coastal GWBs, with chloride values between 0.8 and 2 mg/L. Conversely, fluorides always show very low NBLs. The double PS approach seems more effective for NBLs calculation in coastal aquifers affected by saline contamination, where the use of a fixed Cl limit fails. It may respond to the international needs for a standardized procedure for NBL assessment.

Stock Portfolio Optimization Using a Combined Approach of Multi Objective Grey Wolf Optimizer and Machine Learning Preselection Methods.

The present paper deals with optimizing the stock portfolio of active companies listed on the Tehran Stock Exchange based on the forecast price. This paper is based on a combination of different filtering methods such as optimization of trading rules based on technical analysis (ROC, SMA, EMA, WMA, and MACD at six levels—Very Very Weak (VVW), Very Weak (VW), Weak (W), Strong (S), Very Strong (VS), and Very Very Strong (VVS)), Markov Chains, and Machine Learning (Random Forest and Support Vector Machine) Filter stock exchanges and provide buy signals between 2011 and 2020. In proportion to each combination of filtering methods, a buy signal is issued and based on the mean-variance (M-V) model, the stock portfolio is optimized based on increasing the portfolio return and minimizing the stock portfolio risk. Based on this, out of 480 companies listed on the Tehran Stock Exchange, 85 active companies have been selected and stock portfolio optimization is based on two algorithms, MOGWO and NSGA II. The analysis results show that the use of SVM learning machine leads to minor correlation error than the random forest method. Therefore, this method was used to predict stock prices. Based on the results, it was observed that if the shares of companies are filtered, the risk of transactions decreases, and the return on the stock portfolio increases. Also, if two filtering methods are applied simultaneously, the stock portfolio returns slightly and the risk increases. In the analysis, MOGWO algorithm has obtained 133.13% stock return rate with a risk of 3.346%, while the stock portfolio returns in NSGA II algorithm 107.73, with a risk of 1.459%. Comparison of solution methods shows that the MOGWO algorithm has high efficiency in stock portfolio optimization.

Quantifying nitrate pollution sources and natural background in an equatorial context: A case of the Densu Basin, Ghana

The study assesses the extent of nitrate pollution in the Densu Basin because of itspredominance in agriculture and urbanized vicinities by employing robust techniques for estimating both the natural background and human-induced concentrations. Statistical methods used to estimate these concentrations are the pre-selection method, graphical approach (probability plot), non-parametric approach (kernel density estimation), and parametric approach (Gaussian mixture model). The study shows that the Gaussian mixture model is robust enough in determining the spectral distribution and clustering of the NO3-N concentration in the basin. It estimated the natural background and human induced concentration at 1.7±1.3 mg/L and 9.8±5.6 mg/L, respectively. The results show that the natural background concentration in the basin is more dominant and hence, conducive for drinking. It’s revealed that 26% of anthropogenic sources have leaked into the natural groundwaters of the Basin. The data suggest multiple sources of NO3-N concentration in the groundwater.