Average Cohen's Kappa Research Articles

Automatic sleep-wake classification and Parkinson's disease recognition using multifeature fusion with support vector machine.

Sleep disturbance is a prevalent nonmotor symptom of Parkinson's disease (PD), however, assessing sleep conditions is always time-consuming and labor-intensive. In this study, we performed an automatic sleep-wake state classification and early diagnosis of PD by analyzing the electrocorticography (ECoG) and electromyogram (EMG) signals of both normal and PD rats. The study utilized ECoG power, EMG amplitude, and corticomuscular coherence values extracted from normal and PD rats to construct sleep-wake scoring models based on the support vector machine algorithm. Subsequently, we incorporated feature values that could act as diagnostic markers for PD and then retrained the models, which could encompass the identification of vigilance states and the diagnosis of PD. Features extracted from occipital ECoG signals were more suitable for constructing sleep-wake scoring models than those from frontal ECoG (average Cohen's kappa: 0.73 vs. 0.71). Additionally, after retraining, the new models demonstrated increased sensitivity to PD and accurately determined the sleep-wake states of rats (average Cohen's kappa: 0.79). This study accomplished the precise detection of substantia nigra lesions and the monitoring of sleep-wake states. The integration of circadian rhythm monitoring and disease state assessment has the potential to improve the efficacy of therapeutic strategies considerably.

Classification in Early Fire Detection Using Multi-Sensor Nodes-A Transfer Learning Approach.

Effective early fire detection is crucial for preventing damage to people and buildings, especially in fire-prone historic structures. However, due to the infrequent occurrence of fire events throughout a building's lifespan, real-world data for training models are often sparse. In this study, we applied feature representation transfer and instance transfer in the context of early fire detection using multi-sensor nodes. The goal was to investigate whether training data from a small-scale setup (source domain) can be used to identify various incipient fire scenarios in their early stages within a full-scale test room (target domain). In a first step, we employed Linear Discriminant Analysis (LDA) to create a new feature space solely based on the source domain data and predicted four different fire types (smoldering wood, smoldering cotton, smoldering cable and candle fire) in the target domain with a classification rate up to 69% and a Cohen's Kappa of 0.58. Notably, lower classification performance was observed for sensor node positions close to the wall in the full-scale test room. In a second experiment, we applied the TrAdaBoost algorithm as a common instance transfer technique to adapt the model to the target domain, assuming that sparse information from the target domain is available. Boosting the data from 1% to 30% was utilized for individual sensor node positions in the target domain to adapt the model to the target domain. We found that additional boosting improved the classification performance (average classification rate of 73% and an average Cohen's Kappa of 0.63). However, it was noted that excessively boosting the data could lead to overfitting to a specific sensor node position in the target domain, resulting in a reduction in the overall classification performance.

From Scalp to Ear-EEG: A Generalizable Transfer Learning Model for Automatic Sleep Scoring in Older People.

Sleep monitoring has extensively utilized electroencephalogram (EEG) data collected from the scalp, yielding very large data repositories and well-trained analysis models. Yet, this wealth of data is lacking for emerging, less intrusive modalities, such as ear-EEG. The current study seeks to harness the abundance of open-source scalp EEG datasets by applying models pre-trained on data, either directly or with minimal fine-tuning; this is achieved in the context of effective sleep analysis from ear-EEG data that was recorded using a single in-ear electrode, referenced to the ipsilateral mastoid, and developed in-house as described in our previous work. Unlike previous studies, our research uniquely focuses on an older cohort (17 subjects aged 65-83, mean age 71.8 years, some with health conditions), and employs LightGBM for transfer learning, diverging from previous deep learning approaches. Results show that the initial accuracy of the pre-trained model on ear-EEG was 70.1%, but fine-tuning the model with ear-EEG data improved its classification accuracy to 73.7%. The fine-tuned model exhibited a statistically significant improvement (p < 0.05, dependent t-test) for 10 out of the 13 participants, as reflected by an enhanced average Cohen's kappa score (a statistical measure of inter-rater agreement for categorical items) of 0.639, indicating a stronger agreement between automated and expert classifications of sleep stages. Comparative SHAP value analysis revealed a shift in feature importance for the N3 sleep stage, underscoring the effectiveness of the fine-tuning process. Our findings underscore the potential of fine-tuning pre-trained scalp EEG models on ear-EEG data to enhance classification accuracy, particularly within an older population and using feature-based methods for transfer learning. This approach presents a promising avenue for ear-EEG analysis in sleep studies, offering new insights into the applicability of transfer learning across different populations and computational techniques. An enhanced ear-EEG method could be pivotal in remote monitoring settings, allowing for continuous, non-invasive sleep quality assessment in elderly patients with conditions like dementia or sleep apnea.

The sonographer's and pathologist's perspective of echogenic microfoci in papillary thyroid carcinoma.

Punctate echogenic foci (PEF)/microcalcifications are thought to represent psammoma bodies (PB) in histopathology. However, there are few and contradictory data on this. Different types of sonographic echogenic microfoci (EMF) are seen in papillary thyroid carcinoma (PTC), and their histopathological equivalents are not clearly known. There is also conflicting data on the interobserver agreement between the sonographers on EMF. We prospectively collected US video records of PTC nodules with and without EMF in two large thyroid centers. All video recordings were independently interpreted by three blinded, experienced sonographers. EMF were classified as true microcalcifications (punctate echogenic foci (PEF) ≤1 mm long), linear microechogenities (>1 mm long, posterior acoustic enhancement of the back wall of a microcystic area), comet-tail artifacts/reverberations or linear microechogenities with comet-tail artifacts/reverberations, non-shadowing coarse echogenic foci (>1 mm nonlinear areas) and unclassifiable. Histopathological evaluation was performed by two blinded, qualified pathologists. A total of 114 malignant nodules were included. The average Cohen's kappa (κ) of three sonographers for the EMF presence was 0.775, indicating substantial agreement. A substantial agreement for PEF with 0.658 κ, only fair agreement for other types of EMF with 0.052 to 0.296 κ were detected. EMF were significantly associated with PB and papillae. PEF had an evident relationship with PB in multivariate analysis. There was a strong positive correlation between the amount of PEF and PB (r = 0.634, P < 0.001). PEF in PTC mainly correspond to PB on histopathology. Although observation of EMF varies among sonographers, this inconsistency can be reduced by classifying EMF into subgroups and keeping the term 'PEF' only for true microcalcifications.

Creation of standardized tools to evaluate reporting in health research: Population Reporting Of Gender, Race, Ethnicity & Sex (PROGRES).

Despite increasing diversity in research recruitment, research finding reporting by gender, race, ethnicity, and sex has remained up to the discretion of authors. This study developped and piloted tools to standardize the inclusive reporting of gender, race, ethnicity, and sex in health research. A modified Delphi approach was used to develop standardized tools for the inclusive reporting of gender, race, ethnicity, and sex in health research. Health research, social epidemiology, sociology, and medical anthropology experts from 11 different universities participated in the Delphi process. The tools were pilot tested on 85 health research manuscripts in top health research journals to determine inter-rater reliability of the tools. The tools each spanned five dimensions for both sex and gender as well as race and ethnicity: Author inclusiveness, Participant inclusiveness, Nomenclature reporting, Descriptive reporting, and Outcomes reporting for each subpopulation. The sex and gender tool had a median score of 6 and a range of 1-15 out of 16 possible points. The percent agreement between reviewers piloting the sex and gender tool was 82%. The interrater reliability or average Cohen's Kappa was 0.54 with a standard deviation of 0.33 demonstrating moderate agreement. The race and ethnicity tool had a median score of 1 and a range of 0-15 out of 16 possible points. Race and ethnicity were both reported in only 25.8% of studies evaluated. Most studies that reported race reported only the largest subgroups; White, Black, and Latinx. The percent agreement between reviewers piloting the race and ethnicity tool was 84 and average Cohen's Kappa was 0.61 with a standard deviation of 0.38 demonstrating substantial agreement. While the overall dimension scores were low (indicating low inclusivity), the interrater reliability measures indicated moderate to substantial agreement for the respective tools. Efforts in recruitment alone will not provide more inclusive literature without improving reporting.

Robust Method for Screening Sleep Apnea With Single-Lead ECG Using Deep Residual Network: Evaluation With Open Database and Patch-Type Wearable Device Data

This paper proposes a robust method to screen patients with sleep apnea syndrome (SAS) using a single-lead electrocardiogram (ECG). This method consists of minute-by-minute abnormal breathing detection and apnea-hypopnea index (AHI) estimation. Heartbeat interval and ECG-derived respiration (EDR) are calculated using the single-lead ECG and used to train the models, including ResNet18, ResNet34, and ResNet50. The proposed method, using data from 1232 subjects, was developed with two open datasets and experimental data and evaluated using two additional open datasets and data acquired from an abdomen-attached wearable device (in total, data from 189 subjects). ResNet18 showed the best results, having an average Cohen's kappa coefficient of 0.57, in the abnormal breathing detection. Moreover, SAS patient classification, with 15 as the AHI threshold, yielded an average Cohen's kappa coefficient of 0.71. The results of patient classification were biased toward data from the wearable patch-type device, which may be influenced by different ECG waveforms. The proposed method is tuned with a sample of the data from the device, and the performance result of Cohen's kappa increased from 0.54 to 0.91 for SAS patient classification. Our method, proposed in this paper, achieved equivalent performance results with data recorded using an abdomen-attached wearable device and two open datasets used in previous studies, although the method had not used those data during model training. The proposed method could reduce the development costs of commercial software, as it was developed using open datasets, has robust performance throughout all datasets.

Sleep Monitoring Using Ear-Centered Setups: Investigating the Influence From Electrode Configurations.

Among candidate positions, those in the facial area and around the ears have the benefit of being relatively hairless, and in our view deserve extra attention. In this paper, we seek to determine the limits to sleep monitoring quality within this spatial constraint. We compare 13 different, realistic sensor setups derived from the same data set and analysed with the same pipeline. All setups which include both a lateral and an EOG derivation show similar, state-of-the-art performance, with average Cohen's kappa values of at least 0.80. If large electrode distances are used, positioning is not critical for achieving large sleep-related signal-to-noise-ratio, and hence accurate sleep scoring. We argue that with the current competitive performance of automated staging approaches, there is a need for establishing an improved benchmark beyond current single human rater scoring.

Agricultural burned area detection using an integrated approach utilizing multi spectral instrument based fire and vegetation indices from Sentinel-2 satellite

This study presents a methodology that focuses on detecting agricultural burned areas using Sentinel-2 multispectral data at 10 m. We developed a simple, locally adapted, straightforward approach of multi-index threshold to extract post-winter agricultural burned areas at high resolution for 2019-21. Further, we design a new method for virtual sample collection using already validated fire location data and visual interpretation conditioned using strict selection criteria to improve sample accuracy. Sampling accuracy showed near-perfect agreement with an average Cohen's Kappa value of 0.98. We retrieved monthly ABAs at a resolution of 10 m, and these products were validated against reference burned sample plots identified using visual interpretation of Planet (3m) satellite data. Overall, we found that our method performed well, with an F1 score of 83.63% and low commission (20%) and omission (7%) errors. When compared to global burnt area products, validation accuracy demonstrated an exceptional subpixel scale detecting capability. The study also addresses the complexity of residue burnings and burn signatures’ volatile nature by performing multilevel masking and temporal corrections.•A novel remotely sensed data aided virtual sampling approach to acquire burned and unburned samples.•An integrated method to extract smallholder agricultural burned area using Sentinel-2 multispectral data at a high resolution of 10 m

Gaussian Mixture Model of Ground Filtering Based on Hierarchical Curvature Constraints for Airborne Lidar Point Clouds

This paper proposes a Gaussian mixture model of a ground filtering method based on hierarchical curvature constraints. Firstly, the thin plate spline function is iteratively applied to interpolate the reference surface. Secondly, gradually changing grid size and curvature threshold are used to construct hierarchical constraints. Finally, an adaptive height difference classifier based on the Gaussian mixture model is proposed. Using the latent variables obtained by the expectation-maximization algorithm, the posterior probability of each point is computed. As a result, ground and objects can be marked separately according to the calculated possibility. 15 data samples provided by the International Society for Photogrammetry and Remote Sensing are used to verify the proposed method, which is also compared with eight classical filtering algorithms. Experimental results demonstrate that the average total errors and average Cohen's kappa coefficient of the proposed method are 6.91% and 80.9%, respectively. In general, it has better performance in areas with terrain discontinuities and bridges.

Device to evaluate cleanliness of fiber optic connectors using image processing and neural networks

This work proposes a portable, handheld electronic device, which measures the cleanliness in fiber optic connectors via digital image processing and artificial neural networks. Its purpose is to reduce the evaluation subjectivity in visual inspection done by human experts. Although devices with this purpose already exist, they tend to be cost-prohibitive and do not take advantage of neither image processing nor artificial intelligence to improve their results. The device consists of an optical microscope for fiber optic connector analysis, a digital camera adapter, a reduced-board computer, an image processing algorithm, a neural network algorithm and an LCD screen for equipment operation and results visualization. The image processing algorithm applies grayscale histogram equalization, Gaussian filtering, Canny filtering, Hough transform, region of interest segmentation and obtaining radiometric descriptors as inputs to the neural network. Validation consisted of comparing the results by the proposed device with those obtained by agreeing human experts via visual inspection. Results yield an average Cohen's Kappa of 0.926, which implies a very satisfactory performance by the proposed device.

Concordance analysis of intrapartum cardiotocography between physicians and artificial intelligence-based technique using modified one-dimensional fully convolutional networks

Cardiotocography is a common method of electronic fetal monitoring (EFM) for fetal well-being. Data-driven analyses have shown potential for automated EFM assessment. For this preliminary study, we used a novel artificial intelligence method based on fully convolutional networks (FCNs), with deep learning for EFM evaluation and correct recognition, and its possible role in evaluation of nonreassuring fetal status. We retrospectively collected 3239 EFM labor records from 292 deliveries and neonatal Apgar scores between December 2018 and July 2019 at a single medical center. We analyzed these data using an FCN model and compared the results with clinical practice. The FCN model recognized EFM traces like physicians, with an average Cohen's kappa coefficient of agreement of 0.525 and average area under the receiver operating characteristic curve of 0.892 for six fetal heart rate (FHR) categories. The FCN model showed higher sensitivity for predicting fetal compromise (0.528 vs 0.132) but a higher false-positive rate (0.632 vs 0.012) compared with clinical practice. FCN is a modern technique that may be useful for EFM trace recognition based on its multiconvolutional layered analysis. Our model showed a competitive ability to identify FHR patterns and the potential for evaluation of nonreassuring fetal status.

Apneic Event Estimation only using SpO2 Dynamics in Sleep Apnea Patients.

Nocturnal pulse oximetry has been proposed as a tool for diagnosing sleep apnea. We established criteria in determining previous occurrences of apnea events by extracting quantitative characteristics caused by apnea events over the duration of changes in blood oxygen saturation values in our previous studies. In addition, the apnea-hypopnea index was estimated by regression modeling. In this paper, the algorithm presented in the previous study was applied to the data collected from the sleep medicine center of other hospitals to verify its performance. As a result of applying the algorithm to pulse oximetry data of 15 polysomnographic recordings, the minute-by-minute apneic segment detection exhibited an average accuracy of 87.58% and an average Cohen's kappa coefficient of 0.6327. In addition, the correlation coefficient between the estimated apnea-hypopnea index and the reference was 0.95, and the average absolute error was 5.02 events/h. When the algorithm is evaluated on the data collected by the other sleep medicine center, they still detected semi real-time sleep apnea events and showed meaningful results in estimating apnea-hypopnea index, although their performance was somewhat lower than before. With the recent popularity of devices for mobile healthcare, such as the wearable pulse oximeter, the results of this study are expected to improve the user value of devices by implementing mobile sleep apnea diagnosis and monitoring functions.

Scalable automatic sleep staging in the era of Big Data.

Numerous automatic sleep staging approaches have been proposed to provide an eHealth alternative to the current gold-standard - hypnogram scoring by human experts. However, a majority of such studies exploit data of limited scale, which compromises both the validation and the reproducibility and transferability of such automatic sleep staging systems in real clinical settings. In addition, the computational issues and physical meaningfulness of the analysis are typically neglected, yet affordable computation is a key criterion in Big Data analytics. To this end, we establish a comprehensive analysis framework to rigorously evaluate the feasibility of automatic sleep staging from multiple perspectives, including robustness with respect to the number of training subjects, model complexity, and different classifiers. This is achieved for a large collection of publicly accessible polysomnography (PSG) data, recorded over 515 subjects. The trade-off between affordable computation and satisfactory accuracy is shown to be fulfilled by an extreme learning machine (ELM) classifier, which in conjunction with the physically meaningful hidden Markov model (HMM) of the transition between the different sleep stages (smoothing model) is shown to achieve both fast computation and the highest average Cohen's kappa value of κ = 0.73 (Substantial Agreement). Finally, it is shown that for accurate and robust automatic sleep staging, a combination of structural complexity (multi-scale entropy) and frequency-domain (spectral edge frequency) features is both computationally affordable and physically meaningful.

INTER-RATER AGREEMENT AND VALIDITY OF A TACKLING PERFORMANCE ASSESSMENT SCALE IN YOUTH AMERICAN FOOTBALL

Long term neurologic injury and concussion have been identified as risks from participation in American football. Altering tackling form has been recommended to reduce the risk of neurologic injury caused by head accelerations when tackling. The purpose of this research is to determine the inter-rater agreement and validity of the Qualitative Youth Tackling System (QYTS), a six-item feedback scale to correct tackling form, when utilized by novice and expert raters. Experienced raters will have higher levels of agreement with each other and with motion capture when compared to novice raters. Both novice and experienced raters viewed video of youth athletes (ages 9-13) tackling a dummy in a laboratory setting along. The raters identified successful performance according to a binary rating scale for each component. Analysis of both the raters' agreement with each other and with an objective motion capture measure were completed. Fliess' Kappa measures between all raters were found to be moderate for head placement (k=.48), fair for cervical extension (k=.38), trunk inclination (k=.37), shoulder extension (k=.27) and step length (k=.29), and there was no agreement for pelvic height (k=.-16). When compared to the dichotomized validation measures of each of the five components provided by the motion capture system the average Cohen's Kappa agreement was substantial for pelvic height (k=.63), fair for step length (k=.34), cervical extension (k=.40), trunk inclination (k=.35), and slight for shoulder extension (k=.16). The experienced raters outperformed the novice raters in all categories. The results of this study indicate that skilled raters are better able to identify the movement patterns included in the QYTS when compared to a validation measure as well have higher rates of inter-rater agreement than novice raters. 3b.

Wakefulness evaluation during sleep for healthy subjects and OSA patients using a patch-type device

ObjectivesObstructive sleep apnea (OSA) is a major sleep disorder that causes insufficient sleep, which is linked with daytime fatigue and accidents. Long-term sleep monitoring can provide meaningful information for patients with OSA to prevent and manage their symptoms. Even though various methods have been proposed to objectively measure sleep in ambulatory environments, less reliable information was provided in comparison with standard polysomnography (PSG). Therefore, this paper proposes an algorithm for distinguishing wakefulness from sleep using a patch-type device, which is applicable for both healthy individuals and patients with OSA. MethodsElectrocardiogram (ECG) and 3-axis accelerometer signals were gathered from the single device. Wakefulness was determined with six parallel methods based on information about movement and autonomic nervous activity. The performance evaluation was conducted with five-fold cross validation using the data from 15 subjects with a low respiratory disturbance index (RDI) and 10 subjects with high RDI. In addition, wakefulness information, including total sleep time (TST), sleep efficiency (SE), sleep onset latency (SOL), and wake after sleep onset (WASO), were extracted from the proposed algorithm and compared with those from PSG. ResultsAccording to epoch-by-epoch (30 s) analysis, the performance results of detecting wakefulness were an average Cohen's kappa of 0.60, accuracy of 91.24%, sensitivity of 64.12%, and specificity of 95.73%. Moreover, significant correlations were observed in TST, SE, SOL, and WASO between the proposed algorithm and PSG (p < 0.001). ConclusionsWakefulness-related information was successfully provided using data from the patch-type device. In addition, the performance results of the proposed algorithm for wakefulness detection were competitive with those from previous studies. Therefore, the proposed system could be an appropriate solution for long-term objective sleep monitoring in both healthy individuals and patients with OSA.

Real-Time Automatic Apneic Event Detection Using Nocturnal Pulse Oximetry.

Nocturnal pulse oximetry has been proposed as a simpler alternative to polysomnography in diagnosing sleep apnea. However, existing techniques are limited in terms of inability to provide time information on sleep apnea occurrence. This study aimed to propose a new strategy for near real-time automatic detection of apneic events and reliable estimation of apnea-hypopnea index using nocturnal pulse oximetry. Among 230 polysomnographic recordings with apnea-hypopnea index values ranging from 0 to 86.5 events/h, 138 (60%) and the remaining 92 recordings (40%) were categorized as training and test sets, respectively. By extracting the quantitative characteristics caused by the apneic event for the amount and duration of the change in blood oxygen saturation value, we established the criteria to determine the occurrence of apneic event. Regression modeling was used to estimate the apnea-hypopnea index from the apneic event detection results. The minute-by-minute apneic segment detection exhibited an average accuracy of 91.0% and an average Cohen's kappa coefficient of 0.71. Between the apnea-hypopnea index estimations and reference values, the mean absolute error was 2.30 events/h. The average accuracy of our diagnosis of sleep apnea was 96.7% for apnea-hypopnea index cutoff values of ≥5, 10, 15, and 30 events/h. We developed an effective strategy to detect apneic events by using morphometric characteristics in the fluctuation of blood oxygen saturation values. Our study could be potentially useful in home-based multinight apneic event monitoring for purposes of therapeutic intervention and follow-up study on sleep apnea.

Barriers, Motivators and Perspectives on Minimizing Lunchtime Food Waste of Early Adolescents in the National School Lunch Program

ObjectiveTo determine barriers, motivators and perspectives on minimizing lunchtime food waste of early adolescents in the National School Lunch Program (NSLP) in Hawai'i.MethodsParticipants were early adolescents 9–13 years of age whose parents received SNAP benefits on O'ahu or Hawai'i Island. A semi‐structured interview guide was developed and pilot tested with three participants. An initial codebook was created based on current literature on barriers, motivators and perspectives on minimizing food waste of early adolescents in the US. Interviews were conducted with additional participants (n=27). Following each interview, one researcher (CP) listened to the recorded interview and added new codes to the codebook. Data saturation was achieved when no new codes were identified. All interviews were transcribed and each transcript was coded twice in NVivo using the codebook to ensure no new codes were required. The two coders (CP and KF) then independently collapsed similar codes into higher‐order categories, agreed on a final set, and developed a codebook based on these higher‐order categories. The two coders then coded three transcripts each in NVivo to evaluate interrater reliability and an average Cohen's kappa coefficient of ≥0.8 was achieved. The two coders then coded all transcripts, with an average Cohen's kappa coefficient of ≥0.8. The two coders independently determined key themes and collectively summarized the results.ResultsDislike of the taste of food and school policy were the key barriers to minimizing food waste. Food that did not taste good was left uneaten and thrown into the trash. Participants also reported throwing more food into the trash if they were served versus offered food choices, and if there was no system in place at the school to share uneaten food, compost or give leftovers to animals. Conversely, school policy was a motivator to minimize food waste if schools allowed participants to share, compost or feed their leftovers at lunch to animals. Participants' perspectives on food waste also depended on the taste of the food. They found it acceptable to throw unwanted food into the trash, but unacceptable to discard desirable food. Participants also perceived that their peers did not care about wasting food and that their parents were against food waste.ConclusionImproving the taste of food served at lunch in the NSLP may help reduce food waste and also influence youth to care about wasting food. Schools should regularly monitor and assess the taste of food served and improve the taste if students dislike the food served in the NSLP. Modification of school policy to present students with food choices instead of serving a standard meal and allow students to share, compost or feed leftovers to animals may also reduce waste.Support or Funding InformationThis work was supported by the USDA National Institute of Food and Agriculture, Hatch project HAW2038‐R, managed by the College of Tropical Agriculture and Human Resources.

Detection of fixations and smooth pursuit movements in high-speed eye-tracking data

A novel algorithm for the detection of fixations and smooth pursuit movements in high-speed eye-tracking data is proposed, which uses a three-stage procedure to divide the intersaccadic intervals into a sequence of fixation and smooth pursuit events. The first stage performs a preliminary segmentation while the latter two stages evaluate the characteristics of each such segment and reorganize the preliminary segments into fixations and smooth pursuit events. Five different performance measures are calculated to investigate different aspects of the algorithm's behavior. The algorithm is compared to the current state-of-the-art (I-VDT and the algorithm in [11]), as well as to annotations by two experts. The proposed algorithm performs considerably better (average Cohen's kappa 0.42) than the I-VDT algorithm (average Cohen's kappa 0.20) and the algorithm in [11] (average Cohen's kappa 0.16), when compared to the experts’ annotations.

A Quantitative Analysis of Behavioral Case Linkage for Practice

Behavioral case linkage is the method of identifying crimes committed by the same offender analyzing offender's behavior in crime scenes. Previous researches on behavioral case linkage have not provided enough information of analyzing procedure and predictive validity of quantitative case linkage in practice, especially in case of number of serial offences is more than four. In the present study, we illustrated the procedure of quantitative case linkage for practice, and investigated predictive validity of behavioral case linkage analysis, conducting a simulation experiment analyzing crime scene data from 90 serial rape cases of 18 offenders by hierarchical cluster analysis and non-metric multidimensional scaling (MDS). The results of the study indicated that 66% of the trials were perfectly or satisfactorily performed (average Cohen's kappa=0.78) in a simulation experiment of hierarchical cluster analysis. The results also indicated that MDS could effectively illustrate association between serial rapes by different offenders in most of trials. It is considered that quantitative case linkage based on crime scene behavior has enough predictive validity for helping crime investigations to link a number of serial crimes to one offender and reduce the cost of investigation of serial crimes.

Youth health-risk behavior assessment in Fiji: the reliability of Global School-based Student Health Survey content adapted for ethnic Fijian girls

Objective. The Global School-based Student Health Survey (GSHS) is an assessment for adolescent health-risk behaviors and exposures, supported by the World Health Organization. Although already widely implemented – and intended for youth assessment across diverse ethnic and national contexts – no reliability data have yet been reported for GSHS-based assessment in any ethnicity or country-specific population. This study reports test-retest reliability for GSHS content adapted for a female adolescent ethnic Fijian study sample in Fiji. Design. We adapted and translated GSHS content to assess health-risk behaviors as part of a larger study investigating the impact of social transition on ethnic Fijian secondary schoolgirls in Fiji. In order to evaluate the performance of this measure for our ethnic Fijian study sample (n=523), we examined its test-retest reliability with kappa coefficients, % agreement, and prevalence estimates in a sub-sample (n=81). Reliability among strata defined by topic, age, and language was also examined. Results. Average agreement between test and retest was 77%, and average Cohen's kappa was 0.47. Mean kappas for questions from core modules about alcohol use, tobacco use, and sexual behavior were substantial, and higher than those for modules relating to other risk behaviors. Conclusions. Although test-retest reliability of responses within this country-specific version of GSHS content was substantial in several topical domains for this ethnic Fijian sample, only fair reliability for the module assessing dietary behaviors and other individual items suggests that population-specific psychometric evaluation is essential to interpreting language and country-specific GSHS data.