In today's world, due to the advancement of technology, predicting the students' performance is among the most beneficial and essential research topics. Data Mining is extremely helpful in the field of education, especially for analyzing students' performance. It is a fact that predicting the students' performance has become a severe challenge because of the imbalanced datasets in this field, and there is not any comparison among different resampling methods. This paper attempts to compare various resampling techniques such as Borderline SMOTE, Random Over Sampler, SMOTE, SMOTE-ENN, SVM-SMOTE, and SMOTE-Tomek to handle the imbalanced data problem while predicting students' performance using two different datasets. Moreover, the difference between multiclass and binary classification, and structures of the features are examined. To be able to check the performance of the resampling methods better in solving the imbalanced problem, this paper uses various machine learning classifiers including Random Forest, K-Nearest-Neighbor, Artificial Neural Network, XG-boost, Support Vector Machine (Radial Basis Function), Decision Tree, Logistic Regression, and Naïve Bayes. Furthermore, the Random hold-out and Shuffle 5-fold cross-validation methods are used as model validation techniques. The achieved results using different evaluation metrics indicate that fewer numbers of classes and nominal features will lead models to better performance. Also, classifiers do not perform well with imbalanced data, so solving this problem is necessary. The performance of classifiers is improved using balanced datasets. Additionally, the results of the Friedman test, which is a statistical significance test, confirm that the SVM-SMOTE is more efficient than the other resampling methods. Moreover, The Random Forest classifier has achieved the best result among all other models while using SVM-SMOTE as a resampling method. INDEX TERMS Classification, data mining, educational data mining, imbalanced data problem, machine learning, resampling methods, statistical analysis.

A calibration-free scanned wavelength modulation spectroscopy scheme requiring minimal laser characterization is presented. Species concentration and temperature are retrieved simultaneously from a single fit to a group of 2f/1f-WMS lineshapes acquired in one laser scan. The fitting algorithm includes a novel method to obtain the phase shift between laser intensity and wavelength modulation, and allows for a wavelength-dependent modulation amplitude. The scheme is demonstrated by detection of H(2)O concentration and temperature in atmospheric, premixed CH(4)/air flat flames using a sensor operating near 1.4 µm. The detection sensitivity for H(2)O at 2000 K was 4 × 10(-5) cm(-1) Hz(-1/2), and temperature was determined with a precision of 10 K and absolute accuracy of ~50 K. A parametric study of the dependence of H(2)O and temperature on distance to the burner and total fuel mass flow rate shows good agreement with 1D simulations.

Potassium (K) is an important element related to ash and fine-particle formation in biomass combustion processes. In situ measurements of gaseous atomic potassium, K(g), using robust optical absorption techniques can provide valuable insight into the K chemistry. However, for typical parts per billion K(g) concentrations in biomass flames and reactor gases, the product of atomic line strength and absorption path length can give rise to such high absorbance that the sample becomes opaque around the transition line center. We present a tunable diode laser atomic absorption spectroscopy (TDLAAS) methodology that enables accurate, calibration-free species quantification even under optically thick conditions, given that Beer-Lambert's law is valid. Analyte concentration and collisional line shape broadening are simultaneously determined by a least-squares fit of simulated to measured absorption profiles. Method validation measurements of K(g) concentrations in saturated potassium hydroxide vapor in the temperature range 950-1200 K showed excellent agreement with equilibrium calculations, and a dynamic range from 40 pptv cm to 40 ppmv cm. The applicability of the compact TDLAAS sensor is demonstrated by real-time detection of K(g) concentrations close to biomass pellets during atmospheric combustion in a laboratory reactor.

We present a compact sensor for carbon monoxide (CO) in air and exhaled breath based on a room temperature interband cascade laser (ICL) operating at 4.69 µm, a lowvolume circular multipass cell and wavelength modulation absorption spectroscopy. A fringelimited (1σ) sensitivity of 6.5 × 10 −8 cm −1 Hz -1/2 and a detection limit of 9 ± 5 ppbv at 0.07 s acquisition time are achieved, which constitutes a 25-fold improvement compared to direct absorption spectroscopy. Integration over 10 s increases the precision to 0.6 ppbv. The setup also allows measuring the stable isotope 13 CO in breath. We demonstrate quantification of indoor air CO and real-time detection of CO expirograms from healthy non-smokers and a healthy smoker before and after smoking. Isotope ratio analysis indicates depletion of 13 CO in breath compared to natural abundance.

No abstract