Two-sample Kolmogorov-Smirnov Research Articles

93 Coding Disparity and Specificity During Emergency Department Visits After Transitioning to the Tenth Version of the International Classification of Disease

The purpose of this study was to examine coding disparity and specificity after transitioning from ICD-9 to ICD-10. We sought to evaluate whether the increased granularity inherent in ICD-10 has led to greater specificity in diagnosis and coding. The study took place between 2014 and 2017 before and after the conversion from ICD-9 to ICD-10 within the Veteran’s Health Administration and looked at a national cohort of emergency department visits. Two study groups were looked at. The ICD-9 group looked at the first 6 months of years 2014 and 2015, respectively. The ICD-10 group looked at the first 6 months of 2016 and 2017, respectively. These time periods were selected so as to provide uniformity in the before and after groups as well as to avoid the immediate transition period around October 2015. For each visit, participants were characterized by age, sex, ethnicity, comorbidity, utilization and diagnostic coding disparity and specificity. The Chi-squared test was used to examine differences among characteristics. IRB approval was obtained from the University of Maryland and Baltimore VA Medical Center. The cohort included 2,058,189 emergency department visits by 1,245,586 unique individuals over the study period. When looking at the distribution of codes 5365 unique ICD-9 codes were employed compared with 13,632 ICD-10 codes. A two-sample Kolmogorov-Smirnov (D(2000)=0.255, p<0.001) revealed with both ICD-9 and ICD-10 codes having a very large skew (skewness=3.460, kurtosis=17.005). Similarly, the top 100 ICD-9 codes were used in 58% of encounters while the top 100 ICD-10 codes were used in 50% of encounters. The encounters coded with ICD-10 were not consistently coded with more specificity. Forty-three percent of ICD-9 encounters included multiple ICD codes while only 34% of ICD-10 encounters included multiple codes. Average patient age was 60 years and 90% were male. Among the notable comorbidities were coronary artery disease (18%), diabetes (28%), opioid related disorders (19%) and PTSD (20%.) There were no statistical differences in the demographics within (p=0.318) or between groups (p=0.303), or in the distribution of diagnoses within (p=0.999) or between groups (p-0.905), with 95% confidence intervals and 21 degrees of freedom. The most common ED diagnoses during both the ICD-9 and ICD-10 time periods were back pain followed by bronchitis, chest pain and abdominal pain. When comparing unspecified codes to more specific ones, the percent of unspecified codes for back pain were similar for ICD-9 (32%) and ICD-10 (31%). The percent of specified for bronchitis were also similar for ICD-9 (17%) and ICD-10 (16%) For chest pain, ICD-9 encounters (29%) used more specific codes than ICD-10 (18%). For abdominal pain, ICD-10 encounters (66%) used more specific codes than ICD-9 (35%). We examined coding disparity and specificity using the International Classification of Diseases, after the transition from ICD-9 to ICD-10. The distribution of diagnoses remained unchanged before and after the transition. However, there were fewer encounters with multiple codes with ICD-10. While the encounters after the transition to ICD-10 used more unique codes, providers did not consistently exploit these to create more specific codes. The magnitude of lost revenue has not been studied.

Modeling of time-dependent safety performance using anonymized and aggregated smartphone-based dangerous driving event data

Safety performance functions (SPFs) are generally used to relate exposure to the expected number of crashes aggregated over a long time (e.g. a year) by holding all other risk factors constant, and to identify hotspots that have excessive crashes regardless of different time periods. However, it is highly likely that the relationships of exposure, risk factors and crash occurrence can vary across different times of day. This study aims to establish time-dependent SPFs for urban roads by using large-scale dangerous driving event data captured by smartphones in different times of day. Multivariate conditional autoregressive (MVCAR) models are developed to jointly account for spatial and temporal dependence of crash observations. Results of two-sample Kolmogorov-Smirnov tests affirm the heterogeneity of the safety effects of dangerous driving events in different time periods. Time-dependent hotspots are identified using potential for safety improvement (PSI) metric. The assumption here is that due to the change of traffic conditions and environment across different times of day, safety hotspots for different time periods should be different from each other. According to the results of Wilcoxon signed-rank tests, hotspots identified by times of day are found to be mostly different from each other. The findings of this study provide insights into temporal effects of risk factors and can support the development of time-dependent safety countermeasures. Besides, this study also shows the potential of leveraging anonymized and aggregated dangerous driving data to assess traffic safety issues.

A Sequence-Based Damage Identification Method for Composite Rotors by Applying the Kullback–Leibler Divergence, a Two-Sample Kolmogorov–Smirnov Test and a Statistical Hidden Markov Model

Composite structures undergo a gradual damage evolution from initial inter-fibre cracks to extended damage up to failure. However, most composites could remain in service despite the existence of damage. Prerequisite for a service extension is a reliable and component-specific damage identification. Therefore, a vibration-based damage identification method is presented that takes into consideration the gradual damage behaviour and the resulting changes of the structural dynamic behaviour of composite rotors. These changes are transformed into a sequence of distinct states and used as an input database for three diagnostic models, based on the Kullback–Leibler divergence, the two-sample Kolmogorov–Smirnov test and a statistical hidden Markov model. To identify the present damage state based on the damage-dependent modal properties, a sequence-based diagnostic system has been developed, which estimates the similarity between the present unclassified sequence and obtained sequences of damage-dependent vibration responses. The diagnostic performance evaluation delivers promising results for the further development of the proposed diagnostic method.

Digital impressions in dentistry-accuracy of impression digitalisation by desktop scanners.

To test if the partially digital workflow by digitalisation of the impression reveals a comparable accuracy as the indirect digitalisation of the gypsum cast for 4-unit fixed dental prostheses (FDPs). A titanium model with a tapered full veneer preparation of a molar and premolar was used as analysis model. To receive a virtual three-dimensional reference dataset (REF), it was digitised by industrial computed tomography. Three impression materials were used with individual impression trays (N = 36, n/material = 12): (1) PE (Impregum Penta), (2) PVS-I (Imprint 4 Penta: Super Quick Heavy plus Super Quick Light), and (3) PVS-D (Dimension Penta: H Quick plus L). For partially digital workflow (group IMP), two desktop scanners were used: (1) D810 (3Shape D810) and (2) ZZ (Zirkonzahn S600ARTI). For indirect digitalisation (group CAST), gypsum master casts were manufactured and digitalised using the same desktop scanners. Virtual datasets were superimposed by best fit algorithm, and accuracy was analysed by calculating the Euclidean distances (ED) to the REF (Geomagic Qualify). Statistic was determined (Kruskal-Wallis H test, Mann-Whitney U post hoc analysis, two-sample Kolmogorov-Smirnov test, p < 0.05). ZZ showed for positive deviations superior accuracy for IMP than for CAST. PE and PVS-I showed superior accuracy than PVS-D. D810 showed partially significant better performance with PVS-I and PVS-D than ZZ. The partially digital workflow by digitalisation of the impression can be used for clinical indications of small-span fixed dental prostheses. However, for this indication, the impression material and the desktop scanner are more decisive for the accuracy of virtual model datasets. Despite the rapid advancement of the computer-aided technology for dental therapy purposes, the implementation of this technique is not as fast as the technical development. In order to combine the well-established procedure to use elastomeric materials for a conventional impression and to avoid the drawbacks of casting it by gypsum, the digitalisation of the impression itself by a desktop scanner may be a logical procedure as an access point to the digital workflow. However, there is only limited information about the accuracy of this partially digital workflow by the digitalisation of modern impression materials in comparison to the well-known process of indirect digitalisation of gypsum casts.

Refined asymptotic Kolmogorov-Smirnov tests for the case of finite support

We derive asymptotic critical values for the one-sample and two-sample Kolmogorov-Smirnov tests in the case where the distribution or distributions have finite support of known size. When the number of points in the support is small, these critical values are significantly smaller than those for the standard Kolmogorov-Smirnov tests, thus leading to a major increase in power. We also examine how the asymptotic critical values perform for small sample sizes, finding that they typically lead to conservative inference procedures.

927-P: Use of Mobile-Enabled Reminders Feature Is Associated with Improved Behavioral and Glycemic Outcomes in the Real World

In the real-world, lifestyle and environmental barriers can adversely impact self-management of diabetes. Emerging health technologies such as mobile diabetes management apps can potentially help people with diabetes navigate such barriers with features like reminders for self-monitoring blood glucose (SMBG) checks and medications. While such features are designed to help people with diabetes improve the structure and adherence of SMBG and medication regimens, real-world evidence of their effectiveness have been inconclusive. In the current study, we assessed whether the use of reminders features corresponded with improved behavioral and glycemic outcomes in the real world. We compared the distributions of SMBG check rates, and low (&amp;lt;70 mg/dL), in-range (between 70-180 mg/dL), and high SMBG readings (&amp;gt;180 mg/dL) between users of a mobile diabetes management app who had set up reminders and app users who did not (n=500 in each sample, balanced by age, gender, and app use duration). Two-sample Kolmogorov-Smirnov tests showed that app users with reminders performed more frequent SMBG checks (D=.20, P&amp;lt;.001), had more in-range readings (D=.096, P=.020), and had fewer high readings (D=.086, P=.049) than users without reminders. No difference was observed in the distributions of low readings, though a trend may be present (D=.084, P=.059). These findings indicate that mobile-enabled reminders can contribute to improved SMBG behavior and glycemic outcomes in the real world. As mobile health/digital therapeutic tools (and their ever-growing features/capabilities) continue to evolve, patients and clinicians have reason for optimism that such tools can improve diabetes management and reduce burden. Disclosure R. Abad Juan: Employee; Self; Glooko, Inc. T. Sheng: Employee; Self; Glooko, Inc. S. Babikian: None. M. Greenfield: Stock/Shareholder; Self; Glooko, Inc.

Novel Stopping Criteria for Optimization-Based Microwave Breast Imaging Algorithms.

A discontinuous Galerkin formulation of the Contrast Source Inversion algorithm (DGM-CSI) for microwave breast imaging employing a frequency-cycling reconstruction technique has been modified here to include a set of automated stopping criteria that determine a suitable time to shift imaging frequencies and to globally terminate the reconstruction. Recent studies have explored the use of tissue-dependent geometrical mapping of the well-reconstructed real part to its imaginary part as initial guesses during consecutive frequency hops. This practice was shown to improve resulting 2D images of the dielectric properties of synthetic breast models, but a fixed number of iterations was used to halt DGM-CSI inversions arbitrarily. Herein, a new set of stopping conditions is introduced based on an intelligent statistical analysis of a window of past iterations of data error using the two-sample Kolmogorov-Smirnov (K-S) test. This non-parametric goodness-of-fit test establishes a pattern in the data error distribution, indicating an appropriate time to shift frequencies, or terminate the algorithm. The proposed stopping criteria are shown to improve the efficiency of DGM-CSI while yielding images of equivalent quality to assigning an often liberally overestimated number of iterations per reconstruction.

Fault detection of fuel cell systems based on statistical assessment of impedance data

Accurate online health assessment of fuel cell systems is a key for the timely mitigation and maintenance actions to be taken in order to maximise reliability of operation and useful life span of the cells. The majority of approaches rely on occasional probing of the system with small-amplitude signals around an operating point. The responses are then used to create either a parametric or a non-parametric model of the linearised system dynamics. However, during the probing session, the measurements might be corrupted with random noise and disturbances. Consequently, the evaluated parameters, being points on the impedance curve, parameters of the equivalent circuit models or the distribution of relaxation times, contain some uncertainty. That fact is largely ignored in the state of the art techniques, meaning that only mean value estimates are taken into account in the further analysis. In this paper we use a non-parametric two-sample Kolmogorov-Smirnov test to detect a change in the internal condition by evaluating changes at each frequency point on the Nyquist curve. Moreover, we show that in some cases it is even possible to isolate the fault origin from the pattern of detected changes. The applicability of the approach is demonstrated on the detection of water management faults of an industrial proton exchange membrane fuel cell system.

Statistical Estimation of Perspective Damage Index for Vibration-Based Structural Health Monitoring

Statistical analysis of CCD (correlation coefficient deviation) index for the structural damage detection by vibration-based method was performed. Statistic data set was collected in a full-scale test of a large aircraft component. After special pre-processing the most informative narrow frequency band was selected for obtaining the statistic set of CCD between the frequency response functions of intact and pseudo-damaged states of a structure. The two-sample Kolmogorov-Smirnov hypothesis test was used for estimation of a pseudo-damage effect. The stable response of CCD index to a small damage in large-size structure was demonstrated.

An analytical approach for the simulation of realistic low-dose fluoroscopic images.

The quality of X-ray images plays an important role in computer-assisted interventions. Although learning-based denoising techniques have been shown to be successful in improving the image quality, they often rely on pairs of associated low- and high-dose X-ray images that are usually not possible to acquire at different dose levels in a clinical scenario. Moreover, since data variation is an important requirement for learning-based methods, the use of phantom data alone may not be sufficient. A possibility to address this issue is a realistic simulation of low-dose images from their related high-dose counterparts. We introduce a novel noise simulation method based on an X-ray image formation model. The method makes use of the system parameters associated with low- and high-dose X-ray image acquisitions, such as system gain and electronic noise, to preserve the image noise characteristics of low-dose images. We have compared several corresponding regions of the associated real and simulated low-dose images-obtained from two different imaging systems-visually as well as statistically, using a two-sample Kolmogorov-Smirnov test at 5% significance. In addition to being visually similar, the hypothesis that the corresponding regions-from 80 pairs of real and simulated low-dose regions-belonging to the same distribution has been accepted in 81.43% of the cases. The results suggest that the simulated low-dose images obtained using the proposed method are almost indistinguishable from real low-dose images. Since extensive calibration procedures required in previous methods can be avoided using the proposed approach, it allows an easy adaptation to different X-ray imaging systems. This in turn leads to an increased diversity of the training data for potential learning-based methods.

Cognitive Function in Children with Epilepsy

Introduction: Epilepsy is a worldwide and common child health problem. One of its complications is cognitive impairment that will impact on children’s cognitive development and quality of life. The objective of this study is to find out whether epilepsy affects cognitive function in children.&#x0D; Material and Methods: This was a cross-sectional descriptive study using secondary data from medical records. We collected data from children with epilepsy from 2004 until 2014 that have been actively managed in the Growth and Development Clinic, Sardjito General Hospital, Yogyakarta, Indonesia. Cognitive function referred to low IQ score that was gained from patients’ data. We analyzed the data using Fischer’s exact and two-sample Kolmogorov-Smirnov test methods.&#x0D; Results: Among 40 data samples, there were 19 patients (42.5%) with cognitive impairment (IQ &lt; 70) and 21 patients (57.5%) without cognitive impairment (IQ ≥ 70). There were twenty-five patients examined by the Stanford-Binet IQ test, which found 15 patients with cognitive impairment. In fifteen patients who were examined with the WISC test, we found more children with epilepsy without cognitive impairment (11 patients). Comorbidity with neuro-developmental disorder significantly affected cognitive impairment in children with epilepsy (p value = 0.034, 95% CI 1.08-21.76).&#x0D; Conclusions: More children with epilepsy without cognitive impairment were found. Further studies are needed with larger number of samples utilizing a prospective study design to better determine the association of epilepsy in children and cognitive impairment.

The Potential Utility of Satellite Soil Moisture Retrievals for Detecting Irrigation Patterns in China

Climate change and anthropogenic activities, including agricultural irrigation have significantly altered the global and regional hydrological cycle. However, human-induced modification to the natural environment is not well represented in land surface models (LSMs). In this study, we utilize microwave-based soil moisture products to aid the detection of under-represented irrigation processes throughout China. The satellite retrievals used in this study include passive microwave observations from the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E) and its successor AMSR2, active microwave observations from the Advanced Scatterometer (ASCAT), and the blended multi-sensor soil moisture product from the European Space Agency (i.e., ESA CCI product). We first conducted validations of the three soil moisture retrievals against in-situ observations (collected from the nationwide agro-meteorological network) in irrigated areas in China. It is found that compared to the conventional Spearman’s rank correlation and Pearson correlation coefficients, entropy-based mutual information is more suitable for evaluating soil moisture anomalies induced by irrigation. In general, around 60% of uncertainties in the anomaly of “ground truth” time series can be resolved by soil moisture retrievals, with ASCAT outperforming the others. Following this, the potential utility of soil moisture retrievals in mapping irrigation patterns in China is investigated by examining the difference in probability distribution functions (detected by two-sample Kolmogorov-Smirnov test) between soil moisture retrievals and benchmarks of the numerical model ERA-Interim without considering the irrigation process. Results show that microwave remote sensing provides a promising alternative to detect the under-represented irrigation process against the reference LSM ERA-Interim. Specifically, the highest performance in detecting irrigation intensity is found when using ASCAT in Huang-Huai-Hai Plain, followed by advanced microwave scanning radiometer (AMSR) and ESA CCI. Compared to ASCAT, the irrigation detection capabilities of AMSR exhibit higher discrepancies between descending and ascending orbits, since the soil moisture retrieval algorithm of AMSR is based on surface temperature and, thus, more affected by irrigation practices. This study provides insights into detecting the irrigation extent using microwave-based soil moisture with aid of LSM simulations, which has great implications for numerical model development and agricultural managements across the country.

ENSO modulated groundwater variations in a river basin of Central India

AbstractSeveral previous studies have examined the traceable regional impacts of El Niño Southern Oscillation (ENSO) on groundwater level (GWL) but it remains a question whether the ENSO impacts on groundwater can be established in smaller basins using statistical techniques. The present study attempts to record the ENSO impacts on the groundwater availability in Venna basin, Maharashtra by proposing a combination of statistical and spatial analysis. Utilizing the GWL, the study estimates the spatial variability of GWL, groundwater anomalies, groundwater recharge and discharge using geographic information system (GIS) and quantitative variations in groundwater using statistical techniques. The study also highlights the applicability of the Kolmogorov–Smirnov test in hydrometeorological studies. Analyses reveal the association of deeper GWLs and higher discharge with the El Niño, as opposed to shallower GWLs and higher recharge with La Niña. The two-sample Kolmogorov–Smirnov test confirms the discrepancy in the cumulative distribution of GWL between different ENSO phases. Mann–Kendall, Sen slope and Mann–Whitney tests ascertain the variation of GWL and recharge as well as ENSO impacts in the command area (area irrigated by reservoir using the canal networks) and non-command areas. A significant difference in recharge between El Niño and La Niña is observed in the command as well as in the non-command areas.

Evaluation of Reanalyses over British Columbia. Part I: Daily and Extreme 2-m Temperature

AbstractWeather-station data coverage, quality, and completeness across British Columbia, Canada, degrade outside of population centers and as one goes back in time. This data paucity motivates the search for the best reanalysis to serve as a climatological reference dataset. This study focuses on how well reanalyses represent 2-m temperature (T2M). Systematic error, random error, and two-sample Kolmogorov–Smirnov statistics of daily maximum and minimum T2M are evaluated from the European Centre for Medium-Range Weather Forecasts interim reanalysis (ERA-Interim), the Climate Forecast System Reanalysis (CFSR), the Japanese 55-year Reanalysis (JRA-55), and the Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). Also evaluated are the 2- and 30-yr return levels of T2M, which are estimated by the method of L moments from a fitted generalized extreme value (GEV) distribution. Reanalyses are compared with observations from 57 meteorological stations distributed over the complex terrain of British Columbia from 1980 to 2010. Minimum temperatures are better captured than maximum temperatures by all four reanalyses. JRA-55 and ERA-Interim generally perform better across all metrics. Biases are largely explained by poor reanalysis terrain representation. Statistical stationarity over the 30-yr period is assessed by using Gaussian and GEV distributions fitted with and without time-dependent parameters. It is determined that stationary distributions are sufficient to represent the climate of T2M for this region and time period.

A Speckle Filtering Method Based on Hypothesis Testing for Time-Series SAR Images

To improve the suppression effect for the speckle noise of synthetic aperture radar (SAR) images and the ability of spatiotemporal information preservation of the filtered image without losing the spatial resolution, a novel multitemporal filtering method based on hypothesis testing is proposed in this paper. A framework of a two-step similarity measure strategy is adopted to further enhance the filtering results. Firstly, bi-date analysis using a two-sample Kolmogorov-Smirnov (KS) test is conducted in step 1 to extract homogeneous patches for 3-D patch stacks generation. Subsequently, the similarity between patch stacks is compared by a sliding time-series likelihood ratio (STSLR) test algorithm in step 2, which utilizes the multi-dimensional data structure of the stacks to improve the accuracy of unchanged pixels detection. Finally, the filtered values are obtained by averaging the similar pixels in time-series. The experimental results and analysis of two multitemporal datasets acquired by TerraSAR-X show that the proposed method outperforms the other typical methods with regard to the overall filtering effect, especially in terms of the consistency between the filtered images and the original ones. Furthermore, the performance of the proposed method is also discussed by analyzing the results from step 1 and step 2.

Experimentally informed micromechanical modelling of cement paste: An approach coupling X-ray computed tomography and statistical nanoindentation

This work proposes a method for numerically investigating the fracture mechanism of cement paste at the microscale based on X-ray computed tomography and nanoindentation. For this purpose, greyscale level based digital microstructure was generated by X-ray microcomputed tomography with a resolution of 2 μm/voxel length. In addition, statistics based micromechanical properties (i.e. Young's modulus and hardness) were derived from the grid nanoindentation test which was set to have an interaction volume the same as the resolution of the digital microstructure. A linear relationship between the two probability density functions of greyscale level and local Young's modulus was assumed and verified by the two-sample Kolmogorov-Smirnov (KS) statistic. Based on this assumption, the fracture and deformation of a digital cubic volume with a dimension of 100 μm under uniaxial tension was simulated using a lattice fracture model. In addition, the influence of heterogeneity on fracture response was studied. Furthermore, the proposed method was compared with the results obtained from a traditional approach used previously by the authors in which discrete phases (capillary pore, anhydrous cement clinker, outer and inner hydration products) were considered. The two methods show similar crack patterns and stress-strain responses. The proposed method is regarded more promising as it captures also the gradient of material properties (within the discrete phases) in the cement paste.

Deriving inferential statistics from recurrence plots: A recurrence-based test of differences between sample distributions and its comparison to the two-sample Kolmogorov-Smirnov test.

Recurrence plots (RPs) have proved to be a very versatile tool to analyze temporal dynamics of time series data. However, it has also been conjectured that RPs can be used to model samples of random variables, that is, data that do not contain any temporal dependencies. In the current paper, we show that RPs can indeed be used to mimic nonparametric inferential statistics. Particularly, we use the case of the two-sample Kolmogorov-Smirnov test as a proof-of-concept, showing how such a test can be done based on RPs. Simulations on differences in mean, variance, and shape of two distributions show that the results of the classical two-sample Kolmogorov-Smirnov test and the recurrence-based test for differences in distributions of two independent samples scale well with each other. While the Kolmogorov-Smirnov test seems to be more sensitive in detecting differences in means, the recurrence based test seems to be more sensitive to detect heteroscedasticity and asymmetry. Potential improvements of our approach as well as extensions to tests with individual distributions are discussed.

A monocyte gene expression signature in the early clinical course of Parkinson\u2019s disease

Microglia are the main immune cells of the brain and express a large genetic pattern of genes linked to Parkinson’s disease risk alleles. Monocytes like microglia are myeloid-lineage cells, raising the questions of the extent to which they share gene expression with microglia and whether they are already altered early in the clinical course of the disease. To decipher a monocytic gene expression signature in Parkinson’s disease, we performed RNA-seq and applied the two-sample Kolmogorov-Smirnov test to identify differentially expressed genes between controls and patients with Parkinson's disease and changes in gene expression variability and dysregulation. The gene expression profiles of normal human monocytes and microglia showed a plethora of differentially expressed genes. Additionally, we identified a distinct gene expression pattern of monocytes isolated from Parkinson’s disease patients at an early disease stage compared to controls using the Kolmogorov-Smirnov test. Differentially expressed genes included genes involved in immune activation such as HLA-DQB1, MYD88, REL, and TNF-α. Our data suggest that future studies of distinct leukocyte subsets are warranted to identify possible surrogate biomarkers and may lead to the identification of novel interventions early in the disease course.

Quantification of forest canopy changes caused by spruce budworm defoliation using digital hemispherical imagery

Detection of spruce budworm (Choristoneura fumiferana Clem.) defoliation is critical for forest protection strategies aimed at minimizing losses in growth and mortality. However current aerial and ground survey methods of detecting defoliation are imprecise and subjective, limiting their usefulness. We evaluated the use of hemispherical crown canopy images to quantify annual spruce budworm defoliation, comparing images taken before and after defoliation in each of two years in 75 sample plots in Québec, Canada. Gradient Boosting Machine analysis identified gap fraction change from May-October, gap fraction after defoliation, insecticide spraying, and % balsam fir (Abies balsamea (L.) Mill.) basal area as important explanatory variables of defoliation. Logistic Generalized Linear Model (GLM) and Random Forests (RF) models were trained on a random two-thirds of sample plots combining both years, and defoliation predictions were validated on the remaining one-third of plots. RF predictions consistently resulted in slightly higher correlations and lower root mean squared errors (RMSEs) than GLM predictions. Defoliation models including insecticide spraying, gap fraction change May-October, and % balsam fir basal area had RMSEs of 14–22%, whereas models excluding insecticide spraying had higher RMSEs of 18–24%. Model goodness-of-fit using two-sample Kolmogorov-Smirnov tests indicated that predicted and measured annual defoliation had similar distributions, with the exception of GLM and RF models excluding spray compared to ocular defoliation. Use of hemispherical images to quantify gap fraction change is a feasible, non-destructive, and objective method to assess canopy foliage changes caused by spruce budworm defoliation.

Interval importance index to select relevant ATR-FTIR wavenumber Intervals for falsified drug classification

The commerce of falsified drugs has substantially grown in recent years due to facilitated access to technologies needed for copying authentic pharmaceutical products. Attenuated Total Reflectance coupled with Fourier Transform Infrared (ATR-FTIR) spectroscopy has been successfully employed as an analytical tool to identify falsified products and support legal agents in interrupting illegal operations. ATR-FTIR spectroscopy typically yields datasets comprised of hundreds of highly correlated wavenumbers, which may compromise the performance of classical multivariate techniques used for sample classification. In this paper we propose a new wavenumber interval selection method aimed at selecting regions of spectra that best discriminate samples of seized drugs into two classes, authentic or falsified. The discriminative power of spectra regions is represented by an Interval Importance Index (III) based on the Two-Sample Kolmogorov-Smirnov test statistic, which is a novel proposition of this paper. The III guides an iterative forward approach for wavenumber selection; different data mining techniques are used for sample classification. In 100 replications using the best combination of classification technique and wavenumber intervals, we obtained average 99.87% accurate classifications on a Cialis® dataset, while retaining 12.5% of the authentic wavenumbers, and average 99.43% accurate classifications on a Viagra® dataset, while retaining 23.75% of the authentic wavenumbers. Our proposition was compared with alternative approaches for individual and interval wavenumber selection available in the literature, always leading to more consistent and easier to interpret results.