Multiclass Logistic Regression Research Articles

Multiclass Logistic Regression Classification with PCA for Imbalanced Medical Datasets

Multiclass Logistic Regression Classification with PCA for Imbalanced Medical Datasets

Predicting Air Quality Based on Multiclass Machine Learning Techniques

India is among the most polluted nations with severe environmental implications of pollution increase in several bas cities. For the past few years’ Indian cities have witnessed an alarming drop in the Air Quality (AQ) which is a result of the rapid economic growth. The lives of billions of people are deeply affected by Air Pollution (AP) every year. The economic sectors that source poor urban AQ include transportation, agriculture, construction, forestry, and logging, emitting dangerous gases and particulates such as NO2, NH3, SO3, and Benzene. This research attempts to implement four different classifiers independently on the AQ dataset. The four classifiers that are being implemented are the Multiclass Decision Jungle (MDJ), Multiclass Logistic Regression (MLR), Multiclass Decision Forest (MDF), and Multiclass Neural Network (MNN). This project objective is to identify the most suitable among the four classification models in building the best model for AQ. The performance of a model is judged by accuracy, precision, and recall. Likewise, about other three modes MDF performs best, where it obtained 99.96% accuracies, 98.91% precision, and 99.76% recall.

Machine learning classification of permeable conducting spheres in air and seawater using electromagnetic pulses

Abstract This paper presents machine learning classification on simulated data of permeable conducting spheres in air and seawater irradiated by low frequency electromagnetic pulses. Classification accuracy greater than 90% was achieved. The simulated data were generated using an analytical model of a magnetic dipole in air and seawater placed 1.5–3.5 m above the center of the sphere in 50 cm increments. The spheres had radii of 40 cm and 50 cm and were of permeable materials, such as steel, and non-permeable materials, such as aluminum. A series RL circuit was analytically modeled as the transmitter coil, and an RLC circuit as the receiver coil. Additive white Gaussian noise was added to the simulated data to test the robustness of the machine learning algorithms to noise. Multiple machine learning algorithms were used for classification including a perceptron and multiclass logistic regression, which are linear models, and a neural network, 1D convolutional neural network (CNN), and 2D CNN, which are nonlinear models. Feature maps are plotted for the CNNs and provide explainability of the salient parts of the time signature and spectrogram data used for classification. The pulses investigated, which expand the literature, include a two-sided decaying exponential, Heaviside step-off, triangular, Gaussian, rectangular, modulated Gaussian, raised cosine, and rectangular down-chirp. Propagation effects, including dispersion and frequency dependent attenuation, are encapsulated by the analytical model, which was verified using finite element modeling. The results in this paper show that machine learning methods are a viable alternative to inversion of electromagnetic induction (EMI) data for metallic sphere classification, with the advantage of real-time classification without the use of a physics-based model. The nonlinear machine learning algorithms used in this work were able to accurately classify metallic spheres in seawater even with significant pulse distortion caused by dispersion and frequency dependent attenuation. This paper presents the first effort towards the use of machine learning to classify metallic objects in seawater based on EMI sensing.

Electrodermal activity and heart rate variability–driven personal thermal comfort prediction and explanation

Understanding personal thermal comfort (PTC) is a complex process requiring the prediction of individual satisfaction with a reasonable explanation. However, the explanation goes beyond predictive ability, and the implications of the interaction of physiological and physical factors need to be investigated. Herein, we address these issues with a new approach based on physiological measurement using electrodermal activity (EDA) and heart rate variability (HRV) for prediction and explanation. Feature engineering is proposed to decompose EDA into skin conductance responses and skin conductance levels (SCL) and HRV into low frequency (LF), high frequency (HF), and their ratio (LF/HF ratio). The classification algorithms are used to model the PTC prediction, and multiclass logistic regression (LR) encodes the explanation of predicted results. The predictive effectiveness–based F1-scores of artificial neural networks, K-nearest neighbor, and support vector machine achieve 85 %, 93 %, and 95 %, respectively. Multiclass LR-based odds ratio (OR) investigates the strength of the explanation and highlights that SCL and LF are outstanding explainable features, of which “1” is excluded from the OR sensitivity range. This highlights that the proposed approach based on EDA and HRV helps accomplish predictive and explainable abilities and can be used for future intelligent systems.

A multiclass logistic regression, Markov chain, and cellular automata integration for modelling urban land use change

Cellular Automata (CA) models employ simple transition rules and neighbourhood interactions to determine cells' evolution. Integrating CA with other algorithms is often essential in order to obtain necessary statistical inferences. The integration of Binary Logistic Regression (BLR) and Markov chain (MC) with CA forms common CA integrations. For convenience, urban land use change is often modelled as a binary process (e,g. ‘undeveloped land’ and ‘developed land’), whereas in the real world urban land use change phenomenon could go beyond two processes (e. g. ‘undeveloped land’, ‘sparsely developed land’, ‘semi-highly developed land’, ‘highly developed land’, ‘very highly developed land’, etc.). The BLR-MC-CA hybrid model yields a binary outcome. In order to extend the BLR-MC-CA hybrid model to multivariate problems, we introduce a novel hybrid model, by simply replacing the BLR with the new Multiclass Logistic Regression (MLR) to obtain the new MLR-MC-CA hybrid model. Satellite remote sensing imageries acquired in 1984, 2000, and 2020 are used as input data. Four scenarios of urban land use change are modelled: ‘nonurban region’, ‘highest urban region’, ‘high urban region’, and ‘medium urban region’. These four scenarios are modelled with twelve urban land use drivers: distance to water body, distance to residential areas, distance to industrial and commercial areas, distance to major roads, distance to railway, distance to Lagos Island, distance to international seaport, distance to University of Lagos, income potential, population potential, distance to international airport, and distance to Lagos State University. Urban transitions 1984–2000 and 2000–2020 are used to simulate the urban states in 2000 and 2020 respectively; as well as to obtain the urban state in 2060. The validation of the experiment is largely based on McNemar's test and Receiver Operating Characteristics (ROC) curves. The results of the experiment indicate significant fit between the reference and the predicted data.

Bridging the Intention-Behavior Gap in Organic Food Consumption: Empirical Evidence from China.

The transition to a low-carbon economy emphasizes the importance of green and low-carbon consumption; yet, there is often a discrepancy between consumers' intentions and their actual behavior regarding organic foods. This study aims to identify strategies to bridge this gap. The research model of organic food consumption intention and behavior is constructed, and a structural equation model is used to test the research hypotheses based on a valid sample of 480 residents of Guangdong Province through an online questionnaire survey. Further, the intention‒behavior gap is defined and its determinants are investigated through multiclass logistic regression. Finally, we categorize and forecast the alignment between consumption intentions and behaviors using machine learning algorithms. The results reveal that attitudes, social interactions, and cognitive information play crucial roles in aligning intentions with behaviors. By enhancing social information exchange or improving cognitive understanding, consumers can reduce their intention‒behavior discrepancy. This research offers valuable policy recommendations for fostering green consumption among residents from various perspectives.

Knowledge of Alzheimer’s disease and associated factors among adults in Zhuhai, China: a cross-sectional analysis

BackgroundThis study aimed to assess the public knowledge regarding Alzheimer’s Disease (AD) in Zhuhai, China, focusing on identifying knowledge gaps and the influence of demographic and health factors.MethodsA cross-sectional study was conducted in Zhuhai, China, from October to November 2022. A total of 1986 residents from 18 communities were selected employing stratified multi-stage equi-proportional sampling. Questionnaires covering general information and the Alzheimer’s Disease Knowledge Scale (ADKS) were investigated face-to-face. Ordinal multiclass logistic regression was applied to assess the relationship between AD awareness and demographic and health characteristics.ResultsThe average ADKS score was 18.5 (SD = 3.36) in Zhuhai. The lowest awareness rates were observed in the “Symptoms” and “Caregiving” subdomains of ADKS, with rates of 51.01% and 43.78%, respectively. The correct rates for the 30 ADKS questions ranged from 16.62 to 92.6%, showing a bimodal pattern with clusters around 80% and 20%. Women (OR = 1.203, 95% CI: 1.009–1.435), individuals aged 60 years or older (OR = 2.073, 95% CI: 1.467–2.932), those living in urban areas (OR = 1.361, 95% CI: 1.117–1.662), higher average monthly household income per capita (OR = 1.641, 95% CI: 1.297–2.082), and without any neurological or mental disorders (OR = 1.810, 95% CI: 1.323–2.478) were more likely to have higher levels of awareness about Alzheimer’s disease.ConclusionsAdults in Zhuhai show a limited knowledge of AD, particularly in the ‘Symptoms’ and ‘Caregiving’ subdomains. Upcoming health campaigns must focus on bridging the knowledge gaps in different subdomains of AD, especially among subgroups with lower awareness, as identified in our study.

Investigation of electromagnetic pulse scattering for metallic object classification using machine learning

ABSTRACT This paper presents a metallic object classification method using various electromagnetic pulses. Each electromagnetic pulse irradiated eight metallic objects placed at increasing distances from 10 mm to 40 mm relative to the electromagnetic sensing system. The electromagnetic sensing system consisted of two RL circuits placed in close proximity. Objects were classified using linear (perceptron and multiclass logistic regression) and non-linear (neural network, 1D convolutional neural network (CNN) and 2D CNN) machine learning classifiers. The machine learning classifiers were trained on experimental data collected in an electromagnetically shielded laboratory. A 10-fold cross-validation mean classification accuracy of 99.4 ± 0.3% for the 1D CNN classifier, and 92.9 ± 1.2% for the 2D CNN classifier, was achieved using a rectangular chirp electromagnetic pulse. The rectangular chirp pulse outperformed two-sided decaying exponential, Gaussian, triangular, raised cosine, and rectangular pulses. All pulses had equal energy. While the rectangular chirp performed best overall, other pulses more accurately distinguished between some objects.

Neural networks application based on language features in the classification of complex English textbooks granularity

The surge in modern information has led to a significant increase in text complexity. To meet the needs of various fields for effective information extraction, research on text complexity grading urgently is urgently needed. The study uses the Flesh-Kincaid Grade Level (FKGL) model to extract language features, selects English textbooks as training corpus, and introduces the Graph Convolutional Network of Attention Mechanism (GCN_ATT) model of attention mechanism to construct a text complexity grading model. The research results indicated that in the 10-fold crossover experiment, GCN_ATT’s accuracy, recall, and F1 all reached over 88%. Compared to multi class logistic regression models, GCN_ATT’s various performance indicators were approximately 2% to 3% higher. Meanwhile, GCN_ ATT’s F1 standard deviation decreased by 0.7% and 1.78% compared to the other two models. In addition, GCN_ATT’s fluctuation range of recall and accuracy was less than 20%, a decrease of 12% and 18% compared to the ordered multi classification regression model. Explanation based on GCN_ATT’s text complexity grading has higher accuracy and more stable performance, providing an effective method reference for current text complexity grading problems.

Analysis of Factors Influencing EV Power Battery Recycling Channel Selection: A Study Based on TPB and Multi Class Logistic Regression

New energy electric vehicle (EV) have become the mainstream worldwide, and the power battery recycling industry is facing new opportunities and challenges. This article compares and analyzes the advantages and disadvantages of existing power battery recycling channels. Through a questionnaire survey, a multi class logistic regression model is used to study the impact and correlation of users' gender, age, occupational category, and education level on their choice of recycling channels. The results indicate that there is a positive correlation between education level and understanding of power battery channels. The higher the education level, the more concerned the group is about environmental protection compared to convenience. Finally, based on the TPB, the impact of seven factors on users' choice of recycling channels and corresponding measures were summarized to help the enterprises understand users' inner thoughts, facilitate the establishment of a more comprehensive recycling mechanism, and improve the revenue of power battery recycling.

Unveiling the magic of mega-city block environments: investigating the intriguing mechanisms shaping children's spontaneous play preferences.

This study delves into the spatial preferences of children for play spaces within high-density urban block environments, specifically targeting the area of Baihua Second Road in Shenzhen, China. Recognizing the critical role of play in children's development, and the unique challenges posed by dense urban settings, this research employs multiclass logistic regression models and negative binomial regression models to construct a detailed mathematical analysis of neighborhood spatial elements and children's play space preferences. Data was meticulously gathered through both objective observations of 14 different types of spaces within the block, and subjective assessments via children's responses to a series of environment photos, capturing the essence of over 3,000 child participants' interactions and choices. Key findings reveal a pronounced preference among children for soft facility features and visually appealing spatial experiences, suggesting a nuanced understanding of play space needs beyond traditional playground designs. Notably, the study identifies that while cartoon-style designs in play facilities might increase moderate attractiveness, ordinary designs hold broader appeal, indicating a preference for diversity in play space aesthetics. These insights offer profound implications for urban planners and designers, advocating for a child-centered approach in the creation of urban play environments that prioritize aesthetic diversity, and the integration of natural elements. Moreover, the study situates Baihua Second Road as a paradigmatic case, illustrating the methodology and analytical framework applied in addressing the complex interplay between children's play preferences and urban spatial configurations. By incorporating a comprehensive data-driven analysis, this research contributes significantly to the discourse on child-friendly urban design, offering valuable strategies for cultivating inclusive and engaging urban play spaces for children amidst the constraints of high-density city living.

Masked facial recognition using ensemble convolutional neural network and grey-level co-occurrence matrix

The COVID-19 pandemic proved how face masks became necessary to stop the spread of infection. Due to this, effective identification of people wearing face mask became challenging. Masked facial recognition has significantly increased in accuracy because of developments in convolutional neural networks (CNNs). Small size of the dataset of masked facial images has been a problem in earlier research. As would be expected, this results in poorer accuracy when the model tries to identify faces. In this study, a novel model is proposed with textural feature extraction using grey-level co-occurrence matrix (GLCM) and an ensemble of two pre-trained CNNs DenseNet-121 and VGG-16. Using the minimum redundancy and maximum relevance, the model has improved accuracy by choosing the most important features of the image. The model was trained using in-house dataset that included 38,290 photos of 2,500 people with approximately equal distribution of properly masked, partially masked, and unmasked images. In this, we evaluated the performance of the model on different classifiers multi-class logistic regression (LR) and support vector machine (SVM) with one-vs-rest (OvR) classification and artificial neural network (ANN) and applied a soft voting scheme. The model achieved the highest accuracy of 98.56% at a learning rate of 0.001 on the ANN classifier.

Identifying the personal characteristics of decent work perception for nursing students in China using latent profile analysis.

Given the importance of perceptions of decent workfor nursing students' future career choices, we attempted to determine potential classifications and characteristics of nursing students' perceptions of decent work so that targeted interventions could be developed. A convenience sampleof 1004s- to fourth-year nursing students completed theGeneral Information Questionnaire, Self-Regulatory Fatigue Scale, Occupational Identity Questionnaire, and Decent Work Perceptions Scale in a cross-sectional survey in Heilongjiang Province, China, resulting in 630 valid questionnaires with a valid return rate of 62.75%. Nursing students' perceptions of decent work were defined using descriptive and regression analysis. Latent profile analysis (LPA) identified three subgroups: low perceived decent work group, medium perceived decent work group, and high perceived decent work group, accounting for 4.76%, 69.37%, and 25.87% of the sample, respectively. The results of unordered multiclass logistic regression show that nursing students with relatively low levels of perceived decent work are more likely to have a low professional identity, a lack of respect for nursing seniors, an involuntary choice of nursing major, and a low family income. Different types of nursing students have different perceptions of decent work, and these universities and related departments can use different educational guidance strategies.

A study on the influencing factors of mental health of Chinese garden workers: a cross-sectional study

BackgroundOccupational hazards occur in all walks of life. China’s horticulture industry is undergoing rapid development. However, the mental health of garden workers has not received much attention. This study investigates the mental health status and influencing factors of Chinese garden workers and provides a basis for promoting their mental health and ensuring the healthy development of Chinese horticulture.MethodsA cross-sectional survey of garden workers in Beijing was conducted from 10 July 2021 to 10 October 2021. A total of 3349 valid questionnaires were recovered, with an effective response rate of 95.69%. Descriptive statistical analysis was carried out on the demographic characteristics, job satisfaction, stress, anxiety, and depression of garden workers, and the influencing factors affecting the mental health of Chinese garden workers were found through a t-test, variance analysis, and ordinal multi-class logistic regression analysis.ResultsSurvey respondents were mostly male (54.4%) and under the age of 40 (64.1%). The anxiety and depression symptoms of the garden workers were moderate. Among staff members, 40.2% were in a normal state of stress. Gender, three meals on time, monthly income, and job satisfaction were the factors influencing stress, anxiety, and depression symptoms among garden workers.ConclusionCompared to medical staff and other groups, the stress, anxiety, and depression symptoms of Chinese garden workers are severe. Gender, monthly income, and job satisfaction are important factors affecting their mental health. Managers should continuously improve the working environment of garden workers, provide salaries that match their positions, and improve their job recognition and satisfaction to reduce the impact of negative emotions on personal health.

Machine learning classification of metallic objects using pulse induction electromagnetic data

This paper presents the classification of metallic objects using total and scattered pulse induction electromagnetic data, with a classification accuracy greater than 90%. Machine learning classification is applied to raw electromagnetic induction (EMI) data without the use of a physics-based model. The EMI method is applied to 8 metallic objects placed at increasing distances from 10–55 mm to the EMI sensing system. The EMI sensing system consists of two RL circuits placed in close proximity. Metallic objects are classified using linear algorithms including a perceptron and multiclass logistic regression, and nonlinear algorithms including a neural network, a 1D and 2D convolutional neural network (CNN). EMI data was collected using an experiment in an electromagnetically shielded laboratory. Feature maps are presented that explain the salient components of the EMI data used by the 1D and 2D CNN.

Development of an image-based Random Forest classifier for prediction of surgery duration of laparoscopic sigmoid resections

PurposeSigmoid diverticulitis is a disease with a high socioeconomic burden, accounting for a high number of left-sided colonic resections worldwide. Modern surgical scheduling relies on accurate prediction of operation times to enhance patient care and optimize healthcare resources. This study aims to develop a predictive model for surgery duration in laparoscopic sigmoid resections, based on preoperative CT biometric and demographic patient data.MethodsThis retrospective single-center cohort study included 85 patients who underwent laparoscopic sigmoid resection for diverticular disease. Potentially relevant procedure-specific anatomical parameters recommended by a surgical expert were measured in preoperative CT imaging. After random split into training and test set (75% / 25%) multiclass logistic regression was performed and a Random Forest classifier was trained on CT imaging parameters, patient age, and sex in the training cohort to predict categorized surgery duration. The models were evaluated in the test cohort using established performance metrics including receiver operating characteristics area under the curve (AUROC).ResultsThe Random Forest model achieved a good average AUROC of 0.78. It allowed a very good prediction of long (AUROC = 0.89; specificity 0.71; sensitivity 1.0) and short (AUROC = 0.81; specificity 0.77; sensitivity 0.56) procedures. It clearly outperformed the multiclass logistic regression model (AUROC: average = 0.33; short = 0.31; long = 0.22).ConclusionA Random Forest classifier trained on demographic and CT imaging biometric patient data could predict procedure duration outliers of laparoscopic sigmoid resections. Pending validation in a multicenter study, this approach could potentially improve procedure scheduling in visceral surgery and be scaled to other procedures.

A Multimodal Biometric System for Iris and Face Traits Based on Hybrid Approaches and Score Level Fusion

The increasing need for information security on a worldwide scale has led to the widespread adoption of appropriate rules. Multimodal biometric systems have become an effective way to increase recognition precision, strengthen security guarantees, and reduce the drawbacks of unimodal biometric systems. These systems combine several biometric characteristics and sources by using fusion methods. Through score-level fusion, this work integrates facial and iris recognition techniques to present a multimodal biometric recognition methodology. The Histogram of Oriented Gradients (HOG) descriptor is used in the facial recognition system to extract facial characteristics, while the deep Wavelet Scattering Transform Network (WSTN) is applied in the iris recognition system to extract iris features. Then, for customized recognition classification, the feature vectors from every facial and iris recognition system are fed into a multiclass logistic regression. These systems provide scores, which are then combined via score-level fusion to maximize the efficiency of the human recognition process. The realistic multimodal database known as (MULB) is used to assess the suggested system's performance. The suggested technique exhibits improved performance across several measures, such as precision, recall, accuracy, equal error rate, false acceptance rate, and false rejection rate, as demonstrated by the experimental findings. The face and iris biometric systems have individual accuracy rates of 96.45% and 95.31% respectively. The equal error rates for the face and iris are 1.79% and 2.36% respectively. Simultaneously, the proposed multimodal biometric system attains a markedly enhanced accuracy rate of 100% and an equal error rate as little as 0.26%.

Enhancing functional efficiency in information-extreme machine learning with logistic regression ensembles

The subject matter of the article is the application of supervised machine learning for the task of object class recognition. The goal is enhancing functional efficiency in information-extreme technology (IET) for object class recognition. The tasks to be solved are: to analyse possible ways of increasing the functional efficiency IET approach; implement an ensemble of models that include logistic regression for prioritizing recognition features and an IEI learning algorithm; compare the functional efficiency of the resulting ensemble of models on well-known dataset with classic approach and results of other researchers. The methods: The method is developed within the framework of the functional approach to modelling natural intelligence applied to the problem of object classification. The following results were obtained: The study tries to augment existing IET to support feature prioritization as part of the object class recognition algorithm. The classical information-extreme algorithm treats all input features equivalently important in forming the decisive rule. As a result, the object features with strong correlation are not prioritized by the algorithm's decisive mechanism – resulting in decreasing functional efficiency in exam mode. The proposed approach is solving this problem by applying a two-stage approach. In the first stage the multiclass logistic regression applied to the input training features vectors of objects to be classified – formed the normalized training matrix. To prevent overfitting of the logistic regression, a model the L2(ridge) regularization method was used. On the second stage, the information-extreme method as input takes the result of the first stage. The geometrical parameters of class containers and the control tolerances on the recognition features were considered as the optimization parameters. Conclusions. The proposed approach increases MNIST (Modified National Institute of Standards and Technology) dataset classification accuracy compared with the classic information-extreme method by 26,44%. The proposed approach has a 3.77% lower accuracy compared to neural-like approaches but uses fewer resources in the training phase and allows retraining the model, as well as expanding the dictionary of recognition classes without model retraining.

Volume Expansion Rate Index Reveals the Damage Process of Surrounding Rock: A Machine Learning-Based Effectiveness Evaluation

Energy sustainability and the establishment of the ‘national water network’ are all inseparable from the construction of underground engineering. Monitoring indices for the surrounding rock are vital for the safety management of underground engineering construction, which determines the actual state of the surrounding rock. The complexity of deep underground engineering construction leads to many situations that cannot be predicted and explained by existing experience. Therefore, it is necessary to identify which monitoring index best represents the surrounding rock damage. Currently, there are no advanced and convenient effectiveness evaluation schemes for surrounding rock monitoring information. To fill the technical gap, this study introduces the volume expansion rate (VER) index for surrounding rock and proposes a machine learning (ML)-based evaluation scheme for the effectiveness of monitoring indices. First, six conditions with different in situ stresses are designed, and tunnel excavation monitoring tests are conducted. Second, the surrounding rock damage determination experiments using the ML classification algorithm are performed, and the accuracy matrix and index significance scores are obtained. The evaluation results show that: (1) The multi-class logistic regression algorithm is more suitable for determining surrounding rock damage with high accuracy and more appropriate significance evaluation outcomes. (2) Under the higher in situ stress condition, the tangential stress is more sensitive to the surrounding rock damage. (3) As the in situ stress increases, the significant monitoring indices demonstrate a transition ‘from shallow to deep, from regional damage to point failure’, describing the instability of the surrounding rock and inspiring a new instability criterion for surrounding rock.

Modeling Pedestrian Injury Severity: A Case Study of Using Extreme Gradient Boosting Vs Random Forest in Feature Selection

Walking and bicycling are lauded for their negative net carbon impact and for their health benefits. However, national crash statistics suggest that pedestrians are disproportionately harmed in any vehicle–pedestrian conflict situation. Although automated transportation in the future is anticipated to increase overall safety, multiple incidents involving automated vehicles have been reported recently, indicating that the technology needs more training on real-world scenarios and conflicts. This research is motivated by the need for contextual data and related levels of harm in potential conflict scenarios in mixed traffic and we use a national police reported crash dataset, CRSS, to address this need. Our study uses a new gradient boosting algorithm, XGBoost, to identify important features among a host of seemingly significant variables. We compare the performance of XGBoost with the more frequently used random forest method and find that XGBoost is more reliable and robust for handling an unbalanced and sparse dataset like crash data, and the features extracted are more aligned to findings from previous research on the topic. We also compare feature importance between NASS-GES and CRSS—two national crash databases with different sampling strategies but the same objective—and find that sampling strategy influences selection of feature importance. We further use the features extracted using XGBoost in a multiclass logistic regression to quantify the effect of these features on different levels of pedestrian injury. Our findings indicate that speed limit, light conditions, pre-crash movements, and location of pedestrian are important contributors to crash severity, along with driver distraction and impairment.