Error Predictor Research Articles

Applying Machine Learning to Enhance Runway Safety Through Runway Excursion Risk Mitigation

Risk of runway excursion caused by pilots continuing an unstable approach to landing has been identified by aviation accident investigators as a primary contributing factor in airline landing accidents. The purpose of this research was to develop and test predictive models for unstable approach risk misperception in the National Airspace System using machine learning. The research applied machine learning algorithms to flight recorder data gathered from a fleet of commercial transport aircraft and made available by NASA. Once evidence of unstable approaches was identified and extracted from the flight recorder data, a determination was made whether a rejected landing or continuance to landing was made. Federal Aviation Administration unstable approach criteria were used in the identification of unstable approaches based on flight data variables evaluated at 500 feet above the ground on approach to landing. Six machine learning algorithms were used, including logistic regression, decision tree, gradient boosting, random forest, and support vector machine. The results indicated that the decision tree with three branches produced the best predictive model. This model was able to predict the pilot error of continuing an unstable approach to landing with an accuracy of 98%. Glideslope deviation, selected airspeed, localizer deviation, and flaps not extended were the most important influential predictors of pilot error leading to an unstable approach.

Sampling error in the diagnosis of colorectal cancer is associated with delay to surgery: a retrospective cohort study.

BackgroundAccurate histopathologic diagnosis of colorectal cancer is important for treatment decision-making and timely care. The aim of this study was to measure rates and predictors of sampling errors for biopsy specimens attained at flexible lower gastrointestinal endoscopy, and to determine whether these events lead to a delay in surgical care.MethodsThis is a retrospective observational study of patients who underwent elective resection for colorectal adenocarcinoma between January 2007 and June 2020. Primary outcomes were proportion of incorrectly diagnosed colorectal adenocarcinomas at index endoscopy by histopathology, and time between endoscopy and surgery. Secondary outcomes were predictors of sampling error, and diagnostic yield of repeat endoscopy.ResultsSampling errors occurred in 217/962 (22.6%) flexible endoscopies for colorectal adenocarcinomas. Negative biopsies were associated with a longer median time to surgery (87.6 days, IQR 48.8–180.0) compared to true positive biopsies (64.0 days, IQR 38.0–119.0), p < 0.001. Controlling for lesion location, neoadjuvant therapy, endoscopist specialty, year, and repeat endoscopies, time to surgery remained 1.40-fold longer (p < 0.001) following sampling error. Repeat endoscopy occurred following 62/217 (28.6%) cases of sampling errors, yielding a correct diagnosis of cancer in 38/62 (61.3%) cases. On multivariable analysis, sampling errors were less likely to occur for lesions endoscopists described as suspicious for malignancy (OR 0.12, 95% CI 0.07–0.21) or simple polyps (OR 0.24, 95% CI 0.08–0.70) compared to endoscopically unresectable polyps.ConclusionsColorectal cancers are frequently improperly sampled, which may lead to treatment delays for these patients. When cancer is suspected, surgeons should take care to ensure timely management.

Medication Errors Among Hospitalized Adults in Medical Wards of Nekemte Specialized Hospital, West Ethiopia: A Prospective Observational Study.

PurposeA high prevalence of medication errors in older adults are due to a combination of different factors such as polypharmacy, polymorbidity, enrolment in several disease-management programs, and fragmentation of care that causes medication errors in all age groups. This study aims to assess the incidence and determinants of medication errors among hospitalized adults in medical wards of Nekemte Specialized Hospital (NSH), West Ethiopia.Patients and MethodsA prospective observational study design was conducted at Nekemte Specialized Hospital among hospitalized adults from October 30, 2018 to January 30, 2019. Data were collected by using checklist-guided observation and review of medication order sheets, medication administration records, and patient charts. To identify the independent predictors of medication errors, logistic regression analysis was used. Statistical significance was considered at a p-value <0.05.ResultsA total of 351 patients were included in the present study. The mean age of the patients was 40.67+15.78 years. A total of 813 medication errors were recorded. The most common stage for medication errors was physician ordering 263 (32.4%). In the multivariable analysis, age ≥65 years (AOR: 2.54, 95%CI: 1.12–5.75, p=0.025), hospital stay ≥7 days (AOR: 2.16, 95%CI: 1.17–3.98, p=0.014), number of medication taken (AOR: 1.75, 95%CI: 1.13–2.73, p=0.013) and presence of comorbidity (AOR: 1.57, 95%CI: 1.01–2.46, p=0.049) had shown statistical significance in predicting medication errors.ConclusionMedication errors are common at Nekemte Specialized Hospital with an incidence of 48.3 per 100 orders, 231.6 per 100 admissions, and 433.4 per 1000 patient days. In particular, adults with older age, increased hospital stay, a greater number of medications, and presence of comorbidities were at greater risk for medication errors.

Dysfunctional tissue correlates of unrelated naming errors in acute left hemisphere stroke

ABSTRACT Most naming error lesion-symptom mapping (LSM) studies have focused on semantic and/or phonological errors. Anomic individuals also produce unrelated word errors, which may be linked to semantic or modality-independent lexical deficits. To investigate the neural underpinnings of rarely-studied unrelated errors, we conducted LSM analyses in 100 individuals hospitalised with a left hemisphere stroke who completed imaging protocols and language assessments. We used least absolute shrinkage and selection operator regression to capture relationships between naming errors and dysfunctional brain tissue metrics (regional damage or hypoperfusion in vascular territories) in two groups: participants with and without impaired single-word auditory comprehension. Hypoperfusion—particularly within the parietal lobe—was an important error predictor, especially for the unimpaired group. In both groups, higher unrelated error proportions were associated with primarily ventral stream damage, the language route critical for processing meaning. Nonetheless, brain metrics implicated in unrelated errors were distinct from semantic error correlates.

Relative error prediction: Strong uniform consistency for censoring time series model

This article considers an adaptive method based on the relative error criteria to estimate the regression operator by a kernel smoothing. It is assumed that the variable of interest is subject to random right censoring and that the observations are from a stationary α-mixing process. The uniform almost sure consistency over a compact set with rate where we highlighted the covariance term is established. The simulation study indicates that the proposed approach has better performance in the presence of high level censoring and outliers in data to an existing classical method based on the least squares. An experiment prediction shows the quality of the relative error predictor.

High-Capacity Reversible Data Hiding in Encrypted Images Based on Adaptive Predictor and Compression of Prediction Errors

Reversible data hiding in encrypted images (RDHEI) is a technology which embeds secret data into encrypted images in a reversible way. In this paper, we proposed a novel high-capacity RDHEI method which is based on the compression of prediction errors. Before image encryption, an adaptive linear regression predictor is trained from the original image. Then, the predictor is used to obtain the prediction errors of the pixels in the original image, and the prediction errors are compressed by Huffman coding. The compressed prediction errors are used to vacate additional room with no loss. After image encryption, the vacated room is reserved for data embedding. The receiver can extract the secret data and recover the image with no errors. Compared with existing approaches, the proposed method efficiently improves the embedding capacity.

Prevalence and Risk Factors for Preprocedural Medication Errors in Patients With Atrial Fibrillation and Atrial Flutter.

Perfect adherence to anticoagulant medications is an important aspect of care for patients with atrial fibrillation undergoing cardiac electrophysiology procedures to minimize the risk of stroke. Despite this, adherence remains imperfect as is associated with added cost of additional procedures (e.g., transesophageal echocardiography) and administrative burden. We sought to identify characteristics of such patients and predictors of medication errors at Beth Israel Deaconess Medical Center.

Predictors of refraction prediction error after cataract surgery: a shared parameter model to account for missing post-operative measurements

Predictors of refraction prediction error after cataract surgery: a shared parameter model to account for missing post-operative measurements

Burnout, well‑being, and self‑reported medical errors among physicians.

In recent years, healthcare systems around the world have been subject to profound changes driven by advances in technology, new regulations as well as demographic and societal trends. This volatile and stressful environment has had its consequences for healthcare organizations and their employees. The resulting phenomena of burnout or lowered well‑being may affect key aspects of healthcare delivery on individual, organizational, and financial levels. The objective of this paper was to investigate the level of burnout and well‑being in a group of Polish physicians, its impact on errors, and associations with other factors. This was a cross‑sectional survey using a self‑administered questionnaire with 65 questions and validated tools such as the Maslach Burnout Inventory and the World Health Organization-5 Well‑being Index (WHO‑5). Correlations between variables were investigated using Spearman rank correlation analysis. Univariable and multivariable logistic regression models were used to evaluate predictors of burnout and errors. The study included residents and physicians of various specialties. A total of 125 residents and physicians, mostly internists and cardiologists, participated in the study, out of which 84 (67%) were found to be burned out. The median WHO‑5 score was 40 points. Two‑thirds of all physicians admitted to making an error in the last 3 months, which was strongly associated with burnout. The level of burnout among respondents was high and had numerous negative consequences that may affect the entire healthcare system. The results suggest there is an urgent need to address the problem of burnout among physicians.

Fusion of wildlife tracking and satellite geomagnetic data for the study of animal migration

BackgroundMigratory animals use information from the Earth’s magnetic field on their journeys. Geomagnetic navigation has been observed across many taxa, but how animals use geomagnetic information to find their way is still relatively unknown. Most migration studies use a static representation of geomagnetic field and do not consider its temporal variation. However, short-term temporal perturbations may affect how animals respond - to understand this phenomenon, we need to obtain fine resolution accurate geomagnetic measurements at the location and time of the animal. Satellite geomagnetic measurements provide a potential to create such accurate measurements, yet have not been used yet for exploration of animal migration.MethodsWe develop a new tool for data fusion of satellite geomagnetic data (from the European Space Agency’s Swarm constellation) with animal tracking data using a spatio-temporal interpolation approach. We assess accuracy of the fusion through a comparison with calibrated terrestrial measurements from the International Real-time Magnetic Observatory Network (INTERMAGNET). We fit a generalized linear model (GLM) to assess how the absolute error of annotated geomagnetic intensity varies with interpolation parameters and with the local geomagnetic disturbance.ResultsWe find that the average absolute error of intensity is − 21.6 nT (95% CI [− 22.26555, − 20.96664]), which is at the lower range of the intensity that animals can sense. The main predictor of error is the level of geomagnetic disturbance, given by the Kp index (indicating the presence of a geomagnetic storm). Since storm level disturbances are rare, this means that our tool is suitable for studies of animal geomagnetic navigation. Caution should be taken with data obtained during geomagnetically disturbed days due to rapid and localised changes of the field which may not be adequately captured.ConclusionsBy using our new tool, ecologists will be able to, for the first time, access accurate real-time satellite geomagnetic data at the location and time of each tracked animal, without having to start new tracking studies with specialised magnetic sensors. This opens a new and exciting possibility for large multi-species studies that will search for general migratory responses to geomagnetic cues. The tool therefore has a potential to uncover new knowledge about geomagnetic navigation and help resolve long-standing debates.

Contrastive Analysis and its Implications for Bengali Learners of ESL

Contrastive Analysis, since its introduction as a potential predictor of the influence of the learners’ first language on the second language and the resultant errors, has constantly been subject to criticism on both theoretical and empirical grounds. This once-popular approach of interpreting errors resulting from the interference from the native language met a gradual decline in acceptability as Behaviorism, the theory on which it was based, lost its credibility. Also, empirical findings pointed out its ineffectiveness as the predictor of errors. However, contrastive analysis projects have been carried out in different parts of the world to avoid errors in specific language areas, including pronunciation. This paper aims to shed light on the origin, evolution and significance of the theory, especially in designing instructional materials for the Bengali learners of English as a second language.

Femoral morphology affects postoperative alignment of the lower extremities in hybrid closed-wedge high tibial osteotomy.

This study aimed to determine correction error predictors of lower extremitiy alignment after hybrid closed-wedge high tibial osteotomy (HCWHTO). From 2011 to 2015, 102 knees in 75 patients with medial compartment knee osteoarthritis of the knee and varus knee deformities who underwent HCWHTO were evaluated in this retrospective study with a minimum 2-year follow-up. Preoperative radiological parameters including weight-bearing line ratio (WBL ratio), mechanical leg axis angle, mechanical lateral distal femoral angle (mLDFA), medial proximal tibial angle (MPTA), joint line convergence angle (JLCA), and JLCA under varus or valgus stress were measured. The knees were divided into the following three groups based on the WBL ratio on average 54months after procedure: acceptable (WBL crossing the tibial plateau between 50 and 70%), under-correction (< 50%), and overcorrection (> 70%). A multiple regression analysis was performed to evaluate predictors that could influence postoperative WBL ratio and cut-off values of predictive preoperative factors was analysed. Significant differences in pre- and postoperative WBL ratios and mechanical leg axis angles were found. Preoperative mLDFA in the under-correction group (89.4° ± 2.1°) was significantly higher than that in the overcorrection group (87.4° ± 2.2°) (p = 0.015), while other preoperative parameters revealed no significant differences. The postoperative MPTA in the under-correction group (92.1° ± 3.7°) was significantly lower than that in the other two groups (p = 0.005 and p = 0.0001, respectively), while there was no significant change between the acceptable (94.6° ± 3.4°) and the overcorrection (96.7° ± 3.2°) groups. Multiple regression analysis revealed that mLDFA (β = - 0.316; p = 0.010) had a significant impact on postoperative WBLR. The cut-off value for under-correction of mLDFA was 89.5°and sensitivity and specificity were 53.6 and 71.6%, respectively (p = 0.014). Preoperative mLDFA is significantly associated with postoperative alignment in HCWHTO. While postoperative MPTA is important, disregarding the preoperative mLDFA may bears a high risk of malcorrection. Varus alignment of the knee with larger mLDFA may be an indication for not only HCWHTO, but also for distal femoral osteotomy.

Gender differences in refraction prediction error of five formulas for cataract surgery

ObjectivesTo evaluate gender differences in optical biometry measurements and lens power calculations.MethodsEight thousand four hundred thirty-one eyes of five thousand five hundred nineteen patients who underwent cataract surgery at University of Michigan’s Kellogg Eye Center were included in this retrospective study. Data including age, gender, optical biometry, postoperative refraction, implanted intraocular lens (IOL) power, and IOL formula refraction predictions were gathered and/or calculated utilizing the Sight Outcomes Research Collaborative (SOURCE) database and analyzed.ResultsThere was a statistical difference between every optical biometry measure between genders. Despite lens constant optimization, mean signed prediction errors (SPEs) of modern IOL formulas differed significantly between genders, with predictions skewed more hyperopic for males and myopic for females for all 5 of the modern IOL formulas tested. Optimization of lens constants by gender significantly decreased prediction error for 2 of the 5 modern IOL formulas tested.ConclusionsGender was found to be an independent predictor of refraction prediction error for all 5 formulas studied. Optimization of lens constants by gender can decrease refraction prediction error for certain modern IOL formulas.

Longitudinal predictive ability of mapping algorithms: Secondary analysis of NRG Oncology/RTOG 0415.

60 Background: Mapping algorithms informing economic evaluations are often derived using baseline data from clinical trials. It is unclear if these algorithms can predict health utilities accurately in post-intervention data. Thus, this study examines the longitudinal predictive ability of mapping algorithms derived from baseline trial data and explores the factors associated with prediction errors. Methods: This methodological study utilized data from an international, multicenter, randomized controlled trial of patients with low-risk prostate cancer (PC), conducted by NRG Oncology (NCT00331773). In addition to patient demographic and clinical data, this study utilized PRO data collected at baseline and 6, 12 and 24 months post-intervention. The Expanded Prostate Cancer Index Composite (EPIC) questionnaire measures health-related quality-of-life (HRQoL) and has four domains (urinary, sexual, hormonal, and bowel) and two subdomains per domain (function and bother); EuroQOL-5D-3L (EQ5D) captures health utilities. Ordinary Least Squares (OLS) regression models were used to map EPIC scores to EQ5D utilities in the baseline data through 5-fold cross-validation. Predictive performance was tested in the post-intervention data; predicted and reported utilities were compared using t-tests, and the absolute prediction error was modeled using fixed effects, as a function of baseline demographic and clinical covariates, as well as observed and predicted EQ5D utilities. Results: A total of 267 (199) patients had complete EQ5D and EPIC domain (or subdomain) data at baseline and all subsequent assessments. In the EPIC domain sample, mean ± standard deviation observed EQ5D utility was 0.90±0.13 at baseline, 0.92±0.11 at 6 months, 0.90±0.13 at 12 months and 0.89±0.14 at 24 months. Mean absolute differences (MDs) between reported and predicted were lower for models using EPIC subdomain data compared to EPIC domain data, and generally decreased as the time of assessment increased. The mapping functions over-predicted utilities for patients in perfect health while the prediction errors were increasingly negative for lower reported EQ5D scores. According to the fixed effects model for EPIC domain data, lower observed and predicted baseline EQ5D scores, and time of assessment were significant predictors of the absolute prediction error; for EPIC subdomain data, lower observed and predicted baseline EQ5D scores, hormonal bother and function, and bowel function significantly predicted the absolute prediction error. Conclusions: This study is the first to demonstrate the longitudinal validity of EPIC questionnaire, and builds upon existing research on longitudinal validity of mapping functions. The low MDs in prediction errors in post-intervention data indicate that the mapping functions are sensitive to treatment effect, thereby increasing confidence in their use in economic evaluations in PC.

Combining data assimilation and machine learning to infer unresolved scale parametrization.

In recent years, machine learning (ML) has been proposed to devise data-driven parametrizations of unresolved processes in dynamical numerical models. In most cases, the ML training leverages high-resolution simulations to provide a dense, noiseless target state. Our goal is to go beyond the use of high-resolution simulations and train ML-based parametrization using direct data, in the realistic scenario of noisy and sparse observations. The algorithm proposed in this work is a two-step process. First, data assimilation (DA) techniques are applied to estimate the full state of the system from a truncated model. The unresolved part of the truncated model is viewed as a model error in the DA system. In a second step, ML is used to emulate the unresolved part, a predictor of model error given the state of the system. Finally, the ML-based parametrization model is added to the physical core truncated model to produce a hybrid model. The DA component of the proposed method relies on an ensemble Kalman filter while the ML parametrization is represented by a neural network. The approach is applied to the two-scale Lorenz model and to MAOOAM, a reduced-order coupled ocean-atmosphere model. We show that in both cases, the hybrid model yields forecasts with better skill than the truncated model. Moreover, the attractor of the system is significantly better represented by the hybrid model than by the truncated model. This article is part of the theme issue 'Machine learning for weather and climate modelling'.

National administrative record linkage between specialist community drug and alcohol treatment data (the National Drug Treatment Monitoring System (NDTMS)) and inpatient hospitalisation data (Hospital Episode Statistics (HES)) in England: design, method and evaluation

ObjectivesThe creation and evaluation of a national record linkage between substance misuse treatment, and inpatient hospitalisation data in England.DesignA deterministic record linkage using personal identifiers to link the National Drug...

Modeling Carbohydrate Counting Error in Type 1 Diabetes Management.

Background: The error in estimating meal carbohydrates (CHO) amount is a critical mistake committed by type 1 diabetes (T1D) subjects. The aim of this study is both to investigate which factors, related to meals and subjects, affect the CHO counting error most and to develop a mathematical model of CHO counting error embeddable in T1D patient decision simulators to conduct in silico clinical trials.Methods: A published dataset of 50 T1D adults is used, which includes a patient's CHO count of 692 meals, dietitian's estimates of meal composition (used as reference), and several potential explanatory factors. The CHO counting error is modeled by multiple linear regression, with stepwise variable selection starting from 10 candidate predictors, that is, education level, insulin treatment duration, age, body weight, meal type, CHO, lipid, energy, protein, and fiber content. Inclusion of quadratic and interaction terms is also evaluated.Results: Larger errors correspond to larger meals, and most of the large meals are underestimated. The linear model selects CHO (P < 0.00001), meal type (P < 0.00001), and body weight (P = 0.047), whereas its extended version embeds a quadratic term of CHO (P < 0.00001) and interaction terms of meal type with CHO (P = 0.0001) and fiber amount (P = 0.001). The extended model explains 34.9% of the CHO counting error variance. Comparison with the CHO counting error description previously used in the T1D patient decision simulator shows that the proposed models return more credible realizations.Conclusions: The most important predictors of CHO counting errors are CHO and meal type. The mathematical models proposed improve the description of patients' behavior in the T1D patient decision simulator.

Medication Error and Its Contributing Factors Among Pediatric Patients Diagnosed with Infectious Diseases Admitted to Jimma University Medical Center, Southwest Ethiopia: Prospective Observational Study.

BackgroundMedication errors in pediatric patients are grossly underreported. Pediatric patients are quite susceptible to medication errors. Potential injury by medication error is higher in young children and infants. It results in serious morbidity and mortality. Thus, this study aimed to assess medication error and its contributing factors among pediatric patients diagnosed with infectious diseases admitted to Jimma University Medical Center.MethodsA prospective observational study was conducted among pediatric patients with infectious diseases admitted from April 1 to June 30, 2018. The patient’s written informed consent was obtained after explaining the purpose of the study. The data were collected by structured questionnaire. Data were entered into EpiData version 4.0.2 and then exported to SPSS version 21.0 for analysis. To identify the predictors of medication error, backward logistic regression analysis was done.ResultsFrom a total of the 325 study participants, 136 (41.8%) patients had at least one medication error during their hospital stay. A total of 273 medication errors were identified among 136 patients. Medication errors frequently occurred at prescribing stage 94 (34.4%). The most common types of medication errors were wrong dosing 72 (26.4%) and wrong frequency 47 (17.2%). Presence of disease comorbidity (AOR=1.64, 95%CI=1.01–2.67), being male (AOR=1.79, 95%CI:1.13–2.86) and presence of two infectious diseases (AOR=1.96, 95%CI: 1.20–3.23) and more than three infectious diseases (AOR=2.04, 95%CI: 1.03–4.01) were independent predictors of medication error occurrence.Conclusion and RecommendationMedication errors were common in pediatric patients with infectious diseases in the study area. Presence of comorbidities, being male and the number of infectious diseases were associated with the occurrence of medication errors. Therefore, to reduce medication errors in the study setting, e-prescribing, computerized provider order entry, medication reconciliation, and collaboration of clinical pharmacists with other health professionals are needed.

Personality traits as predictors of cyclist behaviour

Road user behaviour and personality traits are important determinants of driver crash risk. While a great deal of research has been undertaken to understand the relationships between crash involvement, behaviours and personality traits for motor vehicle drivers, comparatively few studies have considered these factors for cyclists. This manuscript presents the findings of a study conducted amongst a sample of six hundred and fifteen (615) Australian cyclists, investigating these issues.The aim of this research was to establish a structure for a cycling behaviour questionnaire applicable to a cohort of Australian cyclists. Using the dimensions identified from the questionnaire, the research investigated the relationship between self-reported crashes, behaviours and personality traits, in order to further develop our understanding of risk factors associated with cycling.Personality traits (agreeableness, extroversion, conscientiousness, neuroticism and openness to experience) were measured using the Big Five Inventory. While cyclist behaviour was measured using a modified version of the cyclist behaviour questionnaire developed by the Dutch national road safety research centre (SWOV).Principal Components Analysis (PCA) was performed on the cycling behaviour questionnaire to identify underlying subscales of behaviour. The PCA identified a two dimension model representing violations (α = 0.74) and errors (α = 0.65), consisting of 16 items from the original 22 item cyclist behaviour questionnaire.Linear regressions for each of the cyclist behaviour factors identified that age was negatively associated with errors and violations, indicating that older cyclists report fewer errors or violations. Similarly, there was a negative association with average weekly kilometres travelled. Gender was a significant predictor of errors, but not violations, with male cyclists reporting fewer errors than females.When considering personality traits, there was a positive association between extroversion and both errors and violations. Significant negative associations were identified for agreeableness and conscientiousness. Neither neuroticism nor openness to experience were associated with the frequency of errors or violations.The research identified that demographics, travel characteristics and personality traits provide insight into engagement in aberrant cycling behaviours and these behaviours are associated with self-reported crash involvement. The research provides insight into behaviours that could be targeted with appropriate education and enforcement strategies.

Sentence contexts and cloze probabilities for Brazilian Portuguese children and adolescents.

PurposeIn this study we investigated a set of 100 sentence contexts and their cloze probabilities to develop a database of linguistic stimuli for Brazilian Portuguese children and adolescents. The study also examined age-related changes on cloze probabilities, and specified the predictor effects of age and cloze probabilities on idiosyncratic responses and errors (semantic, syntactic, and other errors). Finally, the study also aimed to shed light on cultural effects on word generation by comparing Brazilian and Portuguese sentence databases.Method361 typically developing monolingual Brazilian speakers, with ages ranging from 7 to 18 years, participated in the study. The cloze task was composed by 100 sentence contexts, grounded on the European Portuguese database. Responses were classified as valid (correct) or invalid (semantic, syntactic, and other-type errors). Statistical analyses were based on mixed-effects logistic models.ResultsSixty-three sentences met criteria for high cloze probabilities, 30 for medium cloze, and 7 for low cloze. Age was a significant predictor of idiosyncratic responses, semantic and syntactic errors: older participants were less likely to produce idiosyncratic responses, as well as semantic and syntactic errors. Cloze probability values were concordant in the Brazilian and Portuguese databases for 31 out of 49 (83.7%) high-cloze sentences and for 7 low-cloze sentences.ConclusionIn this study we have provided a database with cloze probability values for a set of 100 sentence-final word contexts for Brazilian Portuguese children and adolescents. Results showed that both age and sentence contextual level predicted sentence final word completion. Older participants were more likely to choose more consistently the same final word, with the contextual level of a given sentence also contributing to the final word selection. Age should be controlled for in future studies probing semantic processing with this set of sentences.