Effects of dataset size and interactions on the prediction performance of logistic regression and deep learning models

Evaluation model of urban tourism competitiveness in the context of sustainable development

In the contemporary context marked by globalization and the growing prominence of sustainable development, assessing urban tourism competitiveness has emerged as a crucial research domain. This paper aims to develop a comprehensive model for evaluating city tourism competitiveness, grounded in the principles of sustainable development. The model incorporates factors such as city tourism resources, environmental considerations, economic aspects, and societal factors. This holistic approach seeks to offer valuable insights for the city tourism industry. The study conducts a thorough analysis of current research both domestically and internationally, highlighting gaps and articulating the objectives and significance of the research. Employing a machine learning-based empowerment method, the paper determines the significance of evaluation indices and utilizes the Topsis method for assessing urban tourism competitiveness. Distinguishing itself from traditional evaluation methods, this model integrates the principles of sustainable development throughout the evaluation process, with environmental, social, and economic sustainability serving as pivotal evaluation indicators. Empirical analysis involves the evaluation of tourism competitiveness for select cities, facilitating inter-city comparisons. Results from empirical studies demonstrate the model's effectiveness in evaluating urban tourism competitiveness, providing targeted developmental recommendations for urban tourism.

On the Generalizability of Time-of-Flight Convolutional Neural Networks for Noninvasive Acoustic Measurements

Bulk wave acoustic time-of-flight (ToF) measurements in pipes and closed containers can be hindered by guided waves with similar arrival times propagating in the container wall, especially when a low excitation frequency is used to mitigate sound attenuation from the material. Convolutional neural networks (CNNs) have emerged as a new paradigm for obtaining accurate ToF in non-destructive evaluation (NDE) and have been demonstrated for such complicated conditions. However, the generalizability of ToF-CNNs has not been investigated. In this work, we analyze the generalizability of the ToF-CNN for broader applications, given limited training data. We first investigate the CNN performance with respect to training dataset size and different training data and test data parameters (container dimensions and material properties). Furthermore, we perform a series of tests to understand the distribution of data parameters that need to be incorporated in training for enhanced model generalizability. This is investigated by training the model on a set of small- and large-container datasets regardless of the test data. We observe that the quantity of data partitioned for training must be of a good representation of the entire sets and sufficient to span through the input space. The result of the network also shows that the learning model with the training data on small containers delivers a sufficiently stable result on different feature interactions compared to the learning model with the training data on large containers. To check the robustness of the model, we tested the trained model to predict the ToF of different sound speed mediums, which shows excellent accuracy. Furthermore, to mimic real experimental scenarios, data are augmented by adding noise. We envision that the proposed approach will extend the applications of CNNs for ToF prediction in a broader range.

Automated segmentation and volume prediction in pediatric Wilms' tumor CT using nnu-net

BACKGROUND

Radiologic volumetric evaluation of Wilms' tumor (WT) is an important indicator to guide treatment decisions. However, due to the heterogeneity of the tumors, radiologists have main-guard differences in diagnosis that can lead to misdiagnosis and poor treatment. The aim of this study was to explore whether CT-based outlining of WT foci can be automated using deep learning.

METHODS

We included CT intravenous phase images of 105 patients with WT and double-blind outlining of lesions by two radiologists. Then, we trained an automatic segmentation model using nnUnet. The Dice similarity coefficient (DSC) and 95th percentile Hausdorff distance (HD95) were used to assess the performance. Next, we optimized the automatic segmentation results based on the ratio of the three-dimensional diameter of the lesion to improve the performance of volumetric assessment.

RESULTS

The DSC and HD95 was 0.83 ± 0.22 and 10.50 ± 8.98 mm. The absolute difference and percentage difference in tumor size was 72.27 ± 134.84 cm3 and 21.08% ± 30.46%. After optimization according to our method, it decreased to 40.22 ± 96.06 cm3 and 10.16% ± 9.70%.

CONCLUSION

We introduce a novel method that enhances the accuracy of predicting WT volume by integrating AI automated outlining and 3D tumor diameters. This approach surpasses the accuracy of using AI outcomes alone and has the potential to enhance the clinical evaluation of pediatric patients with WT. By intertwining AI outcomes with clinical data, this method becomes more interpretive and offers promising applications beyond Wilms tumor, extending to other pediatric diseases.

Predicting lncRNA-protein interactions through deep learning framework employing multiple features and random forest algorithm

RNA-protein interaction (RPI) is crucial to the life processes of diverse organisms. Various researchers have identified RPI through long-term and high-cost biological experiments. Although numerous machine learning and deep learning-based methods for predicting RPI currently exist, their robustness and generalizability have significant room for improvement. This study proposes LPI-MFF, an RPI prediction model based on multi-source information fusion, to address these issues. The LPI-MFF employed protein-protein interactions features, sequence features, secondary structure features, and physical and chemical properties as the information sources with the corresponding coding scheme, followed by the random forest algorithm for feature screening. Finally, all information was combined and a classification method based on convolutional neural networks is used. The experimental results of fivefold cross-validation demonstrated that the accuracy of LPI-MFF on RPI1807 and NPInter was 97.60% and 97.67%, respectively. In addition, the accuracy rate on the independent test set RPI1168 was 84.9%, and the accuracy rate on the Mus musculus dataset was 90.91%. Accordingly, LPI-MFF demonstrated greater robustness and generalization than other prevalent RPI prediction methods.

Classification of Cocoa Beans by Analyzing Spectral Measurements Using Machine Learning and Genetic Algorithm

The quality of cocoa beans is crucial in influencing the taste, aroma, and texture of chocolate and consumer satisfaction. High-quality cocoa beans are valued on the international market, benefiting Ivorian producers. Our study uses advanced techniques to evaluate and classify cocoa beans by analyzing spectral measurements, integrating machine learning algorithms, and optimizing parameters through genetic algorithms. The results highlight the critical importance of parameter optimization for optimal performance. Logistic regression, support vector machines (SVM), and random forest algorithms demonstrate a consistent performance. XGBoost shows improvements in the second generation, followed by a slight decrease in the fifth. On the other hand, the performance of AdaBoost is not satisfactory in generations two and five. The results are presented on three levels: first, using all parameters reveals that logistic regression obtains the best performance with a precision of 83.78%. Then, the results of the parameters selected in the second generation still show the logistic regression with the best precision of 84.71%. Finally, the results of the parameters chosen in the second generation place random forest in the lead with a score of 74.12%.

Prediction of drinking water quality with machine learning models: A public health nursing approach

OBJECTIVE

The aim of this study is to use machine learning models to predict drinking water quality from a public health nursing approach.

DESIGN

Machine learning study.

SAMPLE

"Water Quality Dataset" was used in the study. The dataset contains physical and chemical measurements of water quality for 2400 different water bodies. The process consists of four stages: Data processing with Synthetic Minority Oversampling Technique, hyperparameter tuning with 10-fold cross-validation, modeling and comparative analysis. 80% of the dataset is allocated as training data and 20% as test data. ML models logistic regression, K-nearest neighbor, support vector machine, random forest, XGBoost, AdaBoost Classifier, Decision Tree algorithms were used for water quality prediction. Accuracy, precision, recall, F1 score and AUC performance metrics of ML models were compared. To evaluate the performance of the models, 10-fold cross-validation was used and a comparative analysis was performed. The p-values of the models were also compared.

RESULTS

N this study, where drinking water quality was predicted with seven different ML algorithms, it can be said that XGBoost and Random Forest are the best classification models in all performance metrics. There is a significant difference in all ML algorithms according to the p-value. The H0 hypothesis is accepted for these algorithms. According to the H0 hypothesis, there is no difference between actual values and predicted values.

CONCLUSION

In conclusion, the use of ML models in the prediction of drinking water quality can help nurses greatly improve access to clean water, a human right, be more knowledgeable about water quality, and protect the health of individuals.

Variable-complexity machine learning models for large-scale oil spill detection: The case of Persian Gulf

Oil spill is the main cause of marine pollution in the waterbodies with rich oil resources. In this study, we developed and compared the performance of variable-complexity machine-learning models to detect oil spill origin, extent, and movement over large scales. To this end, we trained Support Vector Machine (SVM), Random Forest (RF), and Convolutional Neural Network (CNN) models by using the statistical, geometrical, and textural features of Sentinel-1 SAR data. Our results in the Persian Gulf showed that CNN is superior to RF and SVM classifiers in oil spill detection, as evidenced by the testing accuracy of 95.8 %, 86.0 %, and 78.9 %, respectively. The results suggested utilizing both ascending and descending orbit pass directions to track the movement of oil spill and the underlying transport rate. The proposed methodology enables the detection of probable leaking tankers and platforms, which aids in identifying other sources of oil pollution than tankers and platforms.

Development and validation of machine learning algorithms to predict posthypertensive origin in left ventricular hypertrophy

BACKGROUND

Left ventricular hypertrophy is often associated with hypertension, which is not necessarily the cause of hypertrophy. Non-hypertension-related aetiologies often have a strong impact on patient management, and therefore require a thorough and careful workup. When considering all left ventricular hypertrophies, even the mild ones, the number of patients who need a workup increases drastically. This raises the need for a tool to evaluate the pretest probability of the origin of left ventricular hypertrophy.

AIM

To predict the hypertensive origin of left ventricular hypertrophy using machine learning on first-line clinical, laboratory and echocardiographic variables.

METHODS

We used a retrospective single-centre population of 591 patients with left ventricular hypertrophy, starting at 12mm maximal left ventricular wall thickness. After splitting data in a training and testing set, we trained three different algorithms: decision tree; random forest; and support vector machine. Model performances were validated on the testing set.

RESULTS

All models exhibited good areas under receiver operating characteristic curves: 0.82 (95% confidence interval: 0.77-0.88) for the decision tree; 0.90 (95% confidence interval 0.85-0.94) for the random forest; and 0.90 (95% confidence interval: 0.85-0.94) for the support vector machine. After threshold selection, the last model had the best balance between its specificity of 0.96 (95% confidence interval: 0.91-0.99) and its sensitivity of 0.31 (95% confidence interval: 0.17-0.44). All algorithms relied on similar most influential predictor variables. Online calculators were developed and made publicly available.

CONCLUSIONS

Machine learning models were able to determine the hypertensive origin of left ventricular hypertrophy with good performances. Implementation in clinical practice could reduce the number of aetiological workups needed in patients presenting with left ventricular hypertrophy.

Long-term care insurance purchase decisions of registered nurses: Deep learning versus logistic regression models

OBJECTIVE

The purpose of this study was to use a deep learning model and a traditional statistical regression model to predict the long-term care insurance decisions of registered nurses.

METHODS

We Prospectively surveyed 1,373 registered nurses with a minimum of 2 years of full-time working experience at a large medical center in Taiwan: 615 who already owned long-term care insurance (LTCI), 332 who had no intention to purchase LTCI (group 1), and 426 who intended to purchase LTCI (group 2).

RESULTS

After inverse probability of treatment weighting (IPTW), no statistically significant differences were identified in the study characteristics of the two groups. All the performance indices for the deep neural network (DNN) model were significantly higher than those of the multiple logistic regression (MLR) model (P<0.001). The strongest predictor of an individual's long-term care insurance decision was their risk propensity score, followed by their caregiving responsibilities, whether they live with older adult relatives, their experiences of catastrophic illness, and their openness to experience.

CONCLUSIONS

The DNN model is useful for predicting long-term care insurance decisions. Its prediction accuracy can be increased through training with temporal data collected from registered nurses. Future research can explore designs for two-level or multilevel models that explain the contextual effects of the risk factors on long-term care insurance decisions.

External validity of machine learning-based prognostic scores for cystic fibrosis: A retrospective study using the UK and Canadian registries

Precise and timely referral for lung transplantation is critical for the survival of cystic fibrosis patients with terminal illness. While machine learning (ML) models have been shown to achieve significant improvement in prognostic accuracy over current referral guidelines, the external validity of these models and their resulting referral policies has not been fully investigated. Here, we studied the external validity of machine learning-based prognostic models using annual follow-up data from the UK and Canadian Cystic Fibrosis Registries. Using a state-of-the-art automated ML framework, we derived a model for predicting poor clinical outcomes in patients enrolled in the UK registry, and conducted external validation of the derived model using the Canadian Cystic Fibrosis Registry. In particular, we studied the effect of (1) natural variations in patient characteristics across populations and (2) differences in clinical practice on the external validity of ML-based prognostic scores. Overall, decrease in prognostic accuracy on the external validation set (AUCROC: 0.88, 95% CI 0.88-0.88) was observed compared to the internal validation accuracy (AUCROC: 0.91, 95% CI 0.90-0.92). Based on our ML model, analysis on feature contributions and risk strata revealed that, while external validation of ML models exhibited high precision on average, both factors (1) and (2) can undermine the external validity of ML models in patient subgroups with moderate risk for poor outcomes. A significant boost in prognostic power (F1 score) from 0.33 (95% CI 0.31-0.35) to 0.45 (95% CI 0.45-0.45) was observed in external validation when variations in these subgroups were accounted in our model. Our study highlighted the significance of external validation of ML models for cystic fibrosis prognostication. The uncovered insights on key risk factors and patient subgroups can be used to guide the cross-population adaptation of ML-based models and inspire new research on applying transfer learning methods for fine-tuning ML models to cope with regional variations in clinical care.

Detection and localization of hyperfunctioning parathyroid glands on [18F]fluorocholine PET/ CT using deep learning - model performance and comparison to human experts

BACKGROUND

In the setting of primary hyperparathyroidism (PHPT), [¹⁸F]fluorocholine PET/CT (FCH-PET) has excellent diagnostic performance, with experienced practitioners achieving 97.7% accuracy in localising hyperfunctioning parathyroid tissue (HPTT). Due to the relative triviality of the task for human readers, we explored the performance of deep learning (DL) methods for HPTT detection and localisation on FCH-PET images in the setting of PHPT.

PATIENTS AND METHODS

We used a dataset of 93 subjects with PHPT imaged using FCH-PET, of which 74 subjects had visible HPTT while 19 controls had no visible HPTT on FCH-PET. A conventional Resnet10 as well as a novel mPETResnet10 DL model were trained and tested to detect (present, not present) and localise (upper left, lower left, upper right or lower right) HPTT. Our mPETResnet10 architecture also contained a region-of-interest masking algorithm that we evaluated qualitatively in order to try to explain the model's decision process.

RESULTS

The models detected the presence of HPTT with an accuracy of 83% and determined the quadrant of HPTT with an accuracy of 74%. The DL methods performed statistically worse (p < 0.001) in both tasks compared to human readers, who localise HPTT with the accuracy of 97.7%. The produced region-of-interest mask, while not showing a consistent added value in the qualitative evaluation of model's decision process, had correctly identified the foreground PET signal.

CONCLUSIONS

Our experiment is the first reported use of DL analysis of FCH-PET in PHPT. We have shown that it is possible to utilize DL methods with FCH-PET to detect and localize HPTT. Given our small dataset of 93 subjects, results are nevertheless promising for further research.

On the Automatic Detection and Classification of Skin Cancer Using Deep Transfer Learning

Skin cancer (melanoma and non-melanoma) is one of the most common cancer types and leads to hundreds of thousands of yearly deaths worldwide. It manifests itself through abnormal growth of skin cells. Early diagnosis drastically increases the chances of recovery. Moreover, it may render surgical, radiographic, or chemical therapies unnecessary or lessen their overall usage. Thus, healthcare costs can be reduced. The process of diagnosing skin cancer starts with dermoscopy, which inspects the general shape, size, and color characteristics of skin lesions, and suspected lesions undergo further sampling and lab tests for confirmation. Image-based diagnosis has undergone great advances recently due to the rise of deep learning artificial intelligence. The work in this paper examines the applicability of raw deep transfer learning in classifying images of skin lesions into seven possible categories. Using the HAM1000 dataset of dermoscopy images, a system that accepts these images as input without explicit feature extraction or preprocessing was developed using 13 deep transfer learning models. Extensive evaluation revealed the advantages and shortcomings of such a method. Although some cancer types were correctly classified with high accuracy, the imbalance of the dataset, the small number of images in some categories, and the large number of classes reduced the best overall accuracy to 82.9%.