Stefan Platikanov

Comparison of the variable importance in projection (VIP) and of the selectivity ratio (SR) methods for variable selection and interpretation

This study compares the application of two variable selection methods in partial least squares regression (PLSR), the variable importance in projection (VIP) method and the selectivity ratio (SR) method. For this purpose, three different data sets were analysed: (a) physiochemical water quality parameters related to sensorial data, (b) gas chromatography–mass spectrometry (GC‐MS) chemical (organic compound) profiles from fossil sea sediment samples related to sea surface temperature (SST) changes, and (c) exposed genes of Daphnia magna female samples related to their total offspring production. Correlation coefficients (r), levels of significance (p‐value) and interpretation of the underlying experimental phenomena allowed the discussion about the best approach for variable selection in each case. The comparison of the two variable selection methods in the first water quality data set showed that the SR method is more accurate for sensorial prediction. For the climate data set, when raw total ion current (TIC) GC‐MS chromatograms were considered, variables selected using the VIP method were easier to interpret compared with those selected by the SR method. However, when only some chromatographic peak areas (concentrations) were considered, the SR method was more efficient for prediction, and the VIP method selected the most relevant variables for the interpretation of SST changes. Finally, for the transcriptomic data set, the SR method was found again to be more reliable for prediction purposes. Copyright

Factorial analysis of the trihalomethane formation in the reaction of colloidal, hydrophobic, and transphilic fractions of DOM with free chlorine

Background, aim, and scopeThis study focuses on the factors that affect trihalomethane (THMs) formation when dissolved organic matter (DOM) fractions (colloidal, hydrophobic, and transphilic fractions) in aqueous solutions were disinfected with chlorine.Materials and methodsDOM fractions were isolated and fractionated from filtered lake water and were characterized by elemental analysis. The investigation involved a screening Placket-Burman factorial analysis design of five factors (DOM concentration, chlorine dose, temperature, pH, and bromide concentration) and a Box-Behnken design for a detailed assessment of the three most important factor effects (DOM concentration, chlorine dose, and temperature).ResultsThe results showed that colloidal fraction has a relatively low contribution to THM formation; transphilic fraction was responsible for about 50% of the chloroform generation, and the hydrophobic fraction was the most important to the brominated THM formation.DiscussionWhen colloidal and hydrophobic fraction solutions were disinfected, the most significant factors were the following: higher DOM fraction concentration led to higher THM concentration, an increase of pH corresponded to higher concentration levels of chloroform and reduced bromoform, higher levels of chlorine dose and temperature produced a rise in the total THM formation, especially of the chlorinated THMs; higher bromide concentration generates higher concentrations of brominated THMs. Moreover, linear models were implemented and response surface plots were obtained for the four THM concentrations and their total sum in the disinfection solution as a function of the DOM concentration, chlorine dose, and temperature. Overall, results indicated that THM formation models were very complex due to individual factor effects and significant interactions among the factors.ConclusionsIn order to reduce the concentration of THMs in drinking water, DOM concentrations must be reduced in the water prior to the disinfection. Fractionation of DOM, together with an elemental analysis of the fractions, is important issue in the revealing of the quality and quantity characteristics of DOM. Systematic study composed from DOM fraction investigation and factorial analysis of the responsible parameters in the THM formation reaction can, after an evaluation of the adjustment of the models with the reality, serves well for the evaluation of the spatial and temporal variability in the THM formation in dependence of DOM. However, taking into consideration the natural complexity of DOM, different operations and a strict control of them (like coagulation/flocculation and filtration) has to be used to quantitatively remove DOM from the raw water.Recommendations and perspectivesAssuming that this study represents a local case study, similar experiments can be easily applied and will supply with relevant information every local water treatment plant meeting problems with THM formation. The coagulation/flocculation and the filtration stages are the main mechanisms to remove DOM, particularly the colloidal DOM fraction. With the objective to minimize THMs generation, different unit operation designed to quantitatively remove DOM from water must be optimized.

FTIR-ATR determination of solid non fat (SNF) in raw milk using PLS and SVM chemometric methods

Fourier transform infrared spectroscopy (FTIR) attenuated total reflectance (ATR) spectroscopy, coupled with chemometrics methods have been applied to the fast and non-destructive quantitative determination of solid non fat (SNF) content in raw milk. Partial least squares regression (PLS) and support vector machine (SVM) regression methods were used to model and predict SNF contents in raw milk based on FTIR spectral transmission measurements. Both methods, PLS and SVM, showed good performances in SNF prediction with relative prediction errors in the external validation of between 0.2% and 0.3% depending on the spectral range and regression method. Coefficient of determination of the global fit was always above 0.99. Since, the relative prediction errors were low, it can be concluded that FTIR-ATR with chemometrics can be used for accurate quantitative determinations of SNF contents in raw milk within the investigated calibration range of 79-100g/L. The proposed procedure is fast, non-destructive, simple and easy to implement.

Predicting consumer preferences for mineral composition of bottled and tap water

The overall liking for taste of water was correlated with the mineral composition of selected bottled and tap waters. Sixty-nine untrained volunteers assessed and rated twenty-five different commercial bottled and tap waters from. Water samples were physicochemical characterised by analysing conductivity, pH, total dissolved solids (TDS) and major anions and cations: HCO3-, SO42-, Cl-, NO3-, Ca2+, Mg2+, Na+, and K+. Residual chlorine levels were also analysed in the tap water samples. Globally, volunteers preferred waters rich in calcium bicarbonate and sulfate, rather than in sodium chloride. This study also demonstrated that it was possible to accurately predict the overall liking by a Partial Least Squares regression using either all measured physicochemical parameters or a reduced number of them. These results were in agreement with previously published results using trained panellists.

Linear and non-linear chemometric modeling of THM formation in Barcelona's water treatment plant.

The complex behavior observed for the dependence of trihalomethane formation on forty one water treatment plant (WTP) operational variables is investigated by means of linear and non-linear regression methods, including kernel-partial least squares (K-PLS), and support vector machine regression (SVR). Lower prediction errors of total trihalomethane concentrations (lower than 14% for external validation samples) were obtained when these two methods were applied in comparison to when linear regression methods were applied. A new visualization technique revealed the complex nonlinear relationships among the operational variables and displayed the existing correlations between input variables and the kernel matrix on one side and the support vectors on the other side. Whereas some water treatment plant variables like river water TOC and chloride concentrations, and breakpoint chlorination were not considered to be significant due to the multi-collinear effect in straight linear regression modeling methods, they were now confirmed to be significant using K-PLS and SVR non-linear modeling regression methods, proving the better performance of these methods for the prediction of complex formation of trihalomethanes in water disinfection plants.