F2 Values Research Articles

Comparison of dissolution profiles of brand name and generic drugs of mycophenolate mofetil in the presence of concomitant drugs

Generic drugs have become popular because they reduce medical costs. However, switching from a brand name drug to a generic drug of mycophenolate mofetil (MMF) in immunosuppressive therapy after kidney transplantation resulted in an insufficient increase in its blood concentration in some cases. Immunosuppressive therapy after transplantation typically involves a multidrug strategy. Therefore, changes in efficacy due to switching to a generic drug of MMF may be due to its combination with other drugs. Since the first barrier to the absorption of an orally administered formulation is dissolution, the present study examined the effects of prednisolone (PDN) and tacrolimus (TAC), which are used in immunosuppressive therapy, on the dissolution behavior of MMF. A dissolution experiment was performed in the 1st liquid for dissolution testing (pH 1.2), McIlvaine buffer (pH 4.0), and a 2nd liquid for dissolution testing (pH 6.8) under the single condition of the brand name drug and two generic drugs of MMF as well as under combined conditions with PDN and TAC. Although the initial dissolution rate of MMF differed between the brand name and generic drugs, no significant difference was observed in the final dissolution rate. Furthermore, the f2 value, which evaluates the equivalence of dissolution behavior, was within the standard range. In the 2nd liquid for dissolution testing (pH 6.8), the dissolution rate and ratio slightly decreased under the combined condition with PDN or TAC; however, in comparisons with the single conditions, the f2 value was within the standard range. Therefore, a reduced blood concentration due to switching to a generic drug of MMF may not be due to a change in dissolution behavior.

Artificial intelligence-assisted magnetic resonance imaging technology in the differential diagnosis and prognosis prediction of endometrial cancer.

It aimed to analyze the value of deep learning algorithm combined with magnetic resonance imaging (MRI) in the risk diagnosis and prognosis of endometrial cancer (EC). Based on the deep learning convolutional neural network (CNN) architecture residual network with 101 layers (ResNet-101), spatial attention and channel attention modules were introduced to optimize the model. A retrospective collection of MRI image data from 210 EC patients was used for model segmentation and reconstruction, with 140 cases as the test set and 70 cases as the validation set. The performance was compared with traditional ResNet-101 model, ResNet-101 model based on spatial attention mechanism (SA-ResNet-101), and ResNet-101 model based on channel attention mechanism (CA-ResNet-101), using accuracy (AC), precision (PR), recall (RE), and F1 score as evaluation metrics. Among the 70 cases in the validation set, there were 45 cases of low-risk EC and 25 cases of high-risk EC. Using ROC curve analysis, it was found that the area under the curve (AUC) for the diagnosis of high-risk EC of the proposed model in this article (0.918) was visibly larger as against traditional ResNet-101 (0.613), SA-ResNet-101 (0.760), and CA-ResNet-101 models (0.758). The AC, PR, RE, and F1 values of the proposed model for the diagnosis of EC risk were visibly higher (P < 0.05). In the validation set, postoperative recurrence occurred in 13 cases and did not occur in 57 cases. Using ROC curve analysis, it was found that the AUC for postoperative recurrence prediction of the patients by the proposed model (0.926) was visibly larger as against traditional ResNet-101 (0.620), SA-ResNet-101 (0.729), and CA-ResNet-101 models (0.767). The AC, PR, RE, and F1 values of the proposed model for postoperative recurrence prediction were visibly higher (P < 0.05). The proposed model in this article, assisted by MRI, presented superior performance in diagnosing high-risk EC patients, with higher sensitivity (Sen) and specificity (Spe), and also demonstrated excellent predictive AC in postoperative recurrence prediction.

Optimal design of neural network-based fuzzy predictive control model for recommending educational resources in the context of information technology

Abstract As the information technology develops, educational resource recommendation system has become an indispensable part of the education field. At the same time, people’s demand for personalised educational resources is also growing; therefore, how to accurately predict user demand and provide personalised recommendations has become an important issue. In this study, a fuzzy predictive control model on the grounds of neural network is proposed to optimise the design of educational resource recommendation in the context of information technology. After experimental testing, the model recommendation fit reached 95.16 and 92.91% on the two test datasets, which are significantly higher than the control model. The average F1 values of the proposed model also reached 95.21 and 88.77%, which are higher than the control model. In other control experiments, the proposed model of this study also has a better performance. The relevant outcomes showcase that the predictive performance and recommendation effect of the model can be further enhanced by improving the structure of the neural network and the parameter optimisation method. Meanwhile, the proposed model has high performance in information overload and personalised demand, which offers a useful reference for the optimal design of educational resource recommendation system.

Deep learning models for the prediction of acute postoperative pain in PACU for video‐assisted thoracoscopic surgery

BackgroundPostoperative pain is a prevalent symptom experienced by patients undergoing surgical procedures. This study aims to develop deep learning algorithms for predicting acute postoperative pain using both essential patient details and real-time vital sign data during surgery.MethodsThrough a retrospective observational approach, we utilized Graph Attention Networks (GAT) and graph Transformer Networks (GTN) deep learning algorithms to construct the DoseFormer model while incorporating an attention mechanism. This model employed patient information and intraoperative vital signs obtained during Video-assisted thoracoscopic surgery (VATS) surgery to anticipate postoperative pain. By categorizing the static and dynamic data, the DoseFormer model performed binary classification to predict the likelihood of postoperative acute pain.ResultsA total of 1758 patients were initially included, with 1552 patients after data cleaning. These patients were then divided into training set (n = 931) and testing set (n = 621). In the testing set, the DoseFormer model exhibited significantly higher AUROC (0.98) compared to classical machine learning algorithms. Furthermore, the DoseFormer model displayed a significantly higher F1 value (0.85) in comparison to other classical machine learning algorithms. Notably, the attending anesthesiologists' F1 values (attending: 0.49, fellow: 0.43, Resident: 0.16) were significantly lower than those of the DoseFormer model in predicting acute postoperative pain.ConclusionsDeep learning model can predict postoperative acute pain events based on patients' basic information and intraoperative vital signs.

Zygosity Effects on Human Voice: Fundamental Frequency Analysis of Brazilian Twins' Speech.

Voice production can be influenced by interindividual variations related to genetic, physiological, behavioral, and several environmental factors. Here we examined the effect of zygosity on speaking fundamental frequency (F0) statistical descriptors. Our aims were: (1) to determine whether the genetic similarity between monozygotic (MZ) and dizygotic (DZ) twins affects F0 characteristics, and (2) to quantify the contribution of genetic factors to these characteristics. The study involved 79 same-sex twin pairs of Brazilian Portuguese speakers, comprising 65 MZ and 14 DZ twins, aged 18 to 66 years (31.7 ± 11.6 years), with 21 male and 58 female pairs. Participants were recorded while uttering a greeting phrase and the Brazilian Portuguese version of the 'Happy Birthday to You' song. Speech segments were analyzed using Praat free software, and F0 measures were automatically extracted in both Hertz and semitone scales. Statistical descriptors, including centrality, dispersion, and extreme values of F0 were examined, and the ACE model (i.e., total genetic effects, A; shared environmental influences, C; and nonshared environmental influences, E) was employed to estimate the additive effect;ts of monozygosity. As anticipated, we observed a zygosity effect on several F0 parameters, with more similarity between MZ twins compared to DZ twins. We discuss the genetic influences on F0 parameters and the absence of a monozygosity effect in two of them. Additionally, we briefly address potential biases associated with the selected measurement scale for statistical modeling. Finally, we explore the influence of genetic factors on F0 patterns, as well as environmental, life history and linguistic factors, particularly concerning F0 variation in speech.

Computer-aided engineering: Quantification of the heating non-uniformity and distribution of the thermal load occurring during continuous ohmic and conventional thermal food sterilization

The characterization of continuous thermal processing (CTP) is a crucial aspect in the design and selection of technologies for the production of safe products with optimal quality retention after the thermal stress. Conventional methods for CTP characterization are constrained in their capacity to comprehensively capture the complex dynamics of fluid flow through pipelines and the different heating principles that contribute to local temperature variations. These methods rely on isolated local temperature measurements, which fail to account for the intricate interactions between temperature, heat transfer phenomena and fluid dynamics. In view of the aforementioned limitations, this study presents the implementation of a computational fluid dynamic digital model as a toolbox for the characterization and comparison of a conventional heating (CH) ultra-high temperature (UHT) sterilization process and ohmic heating (OH) UHT sterilization. This serves as a tool for an extensive and accurate comparison of the two processes. The model enabled the estimation of each technology's thermal load through the F0 value, thereby providing a more comprehensive assessment than local temperature measurements alone. Furthermore, this approach accounted for the flow behavior throughout the sterilization process. This strategy demonstrated that 2.5 % of the food product is exposed to an up to 75.4 times the average thermal load, whereas for OH treatments only 0.4 % of the product was exposed up to 5.1 times the average thermal load. This is due to the volumetric heating feature of OH, which leads to a 54.7 % reduction in the cooking grade of the product. Furthermore, the computer-aided comparison revealed no statistically significant difference (p-value of 0.6) between the two technologies in their capacity to inactivate Geobacillus stearothermophilus spores in terms of thermal load. This study highlights the importance of computer-aided engineering methodologies for the technological assessment of food sterilization processes prior to industrial transfer.

Penentuan Kecukupan Panas serta Evaluasi Sensori Produk Opor Ayam Steril PT XYZ

Chicken opor, a seasoned dish made with chicken, spices, and coconut milk, has the potential to be developed into a practical and shelf-stable food product using proper preservation technology. Besides utilizing retort pouch packaging, a key factor in processing sterile chicken opor is ensuring an effective sterilization process. This study aimed to evaluate the thermal adequacy of the sterilization process for chicken opor produced by PT XYZ and to assess consumer acceptance of its sensory attributes. The product was packaged in 300 g retort pouches, with 44 units filling the sterilization basket. Heat distribution was tested using 30 temperature sensors, and heat penetration was assessed with 15 sensors placed at various locations in a 50 L pressure cooker. Sensory evaluation was conducted with 75 panelists using a hedonic rating scale. The results showed that heat distribution within the pressure cooker was uniform. The thermal adequacy, necessary to meet the commercial sterility standard, was achieved with a 70 minutes sterilization process after reaching the maximum temperature of 110 °C, yielding a minimum F0 value of 4.8 minutes. The hedonic rating test indicated moderate to high acceptance of PT XYZ's sterilized chicken opor among the panelists.

The real-time shadow detection of the PV module by computer vision based on histogram matching and gamma transformation method

Solar energy plays an important role in renewable energy generation, with the advantages of low pollution, easy installation, and relatively easy access. However, photovoltaic (PV) modules are susceptible to cause localized shading from external factors such as leaves in the canopy, surrounding buildings, etc., which would affect power generation efficiency and even pose safety risks. Existing methods cannot perform well in real-time conditions. This paper proposes a real-time shading monitoring method for the PV module based on computer vision. The gamma transform and histogram matching were adopted to enhance key features and adjust the global gamut strength distribution in the image of the PV module; then the gray-level slicing method finished the segmentation to detect the shadow from the video. All processing can be realized in the real-time monitor camera and the detection results can be displayed on the HMI in PC with high efficiency and low cost. According to tests in the practical complex environment, the method can have enough detection performance and high real-time performance with an accuracy of 0.98, and the F0.5 and F2 values are 0.87 and 0.85, respectively. The metrics of the proposed method are higher than those of the existing Canny detection method, the Random Forest detection method, and the CNN detection method. In addition, the average time required by the proposed method to process a frame is 0.721 s. In addition, the average time required by the method to process an image frame is 0.721 s, which has good real-time performance.

Development of machine learning algorithm for loop-mediated isothermal amplification including influence of temperature

To predict amplification in the Loop-Mediated Isothermal Amplification (LAMP) method, a binary model has been developed. This model assesses whether the combination of partial template binding facilitated by multiple primers aligns with a specific signature indicating amplification. For this model, 1512 settings were adjusted, encompassing variations in the extraction oligomer position from the primer, the oligomer length, the number of allowed mismatched bases, and the stringency of the homology position. The optimal set was determined based on the F1 value at a reaction temperature of 60°C, resulting in a value of 0.58. To implement machine learning by amalgamating multiple models, duplicate models were eliminated through cluster analysis. Machine learning using AutoML was performed on 108 distinct models. Consequently, the F1 values were 0.7119 at a reaction temperature of 60°C and 0.7210 at 68°C. The weak correlation between feature importance and the F1 value of the model alone suggests that machine learning utilizing this model incorporates numerous models to enhance prediction accuracy.

Determination of bioequivalence between generic and reference drugs using laser-induced breakdown spectroscopy

BackgroundIn vitro bioequivalence studies are strictly limited to the comparison of dissolution performance to a reference drug. These studies are performed without considering the chemical similarity between the generic and reference drug formulations. This work has focused on developing a groundbreaking method based on the laser-induced breakdown spectroscopy (LIBS) technique for the in vitro bioequivalence determination of immediate-release solid oral dosage form generic drugs and as an alternative method for establishing the biowaiver of in vivo generic drug studies. ResultsThe novel LIBS-based methodology to determine in vitro bioequivalence is fast, easy to perform, and can be carried out without the requirement of tedious and complicated sample pre-treatment, nor expensive instrumentals and reagents, almost directly on the drug samples. Furthermore, the proposed methodology demonstrated that it is enough to identify the spectrochemical similarity of the formulation between generic drugs to a reference drug through the chemometric study of their LIBS spectra, based on the determination of the differentiation and similarity factors, f1 and f2, respectively, used in the pharmaceutical industry in this purpose. After analysing their LIBS spectra, the generic drugs selected for this work have all been shown to be in vitro bioequivalent, given their f1 values of less than 15 and f2 values greater than 50, according to the technical regulations on which the American and European medicines agencies are based for the approval of registration for generic immediate-release solid oral dosage form drugs. This has been evidenced even for drugs from Class III and Class IV of the biopharmaceutical classification system, whose active principle nominal concentration is very low as 0.1 and 0.25 mg/tablet, respectively. Significancefor the first time the LIBS technique has been successfully used in an advanced application for the pharmaceutical industry. The proposed method constitutes a reliable and specialized methodology for the establishment of formulation similarity between two drugs, without the requirement of separate identification of each of their components, which is a new and potential tool to determine the in vitro bioequivalence for generic immediate-release solid oral dosage form drugs.

A transfer learning based patent transfer prediction model: Discussion about the condition of China and U.S.

In the same patent classification, there are significant differences in the patent transfer quantity for different countries. For the countries with small patent transfer quantities, the performance of the patent transfer prediction is worse than that of high-transfer countries. To improve the performance of the patent prediction model learned for the low-transfer countries, we propose a transfer learning based patent transfer prediction model, for which the high-transfer countries are the source domain and the low-transfer countries are the target domain. In the model, the textual feature of each patent text is embedded using BERT-CNN with the parameters fine-tuned on the target domain from the source domain. The textual features are optimized to minimize the domain loss of Maximum Mean Discrepancy between source and target domain. What is more, we extract graph feature vectors by embedding the patent node in a three-layer heterogeneous network which consists of patent nodes, International Patent Classification nodes, consumer nodes, consumer location nodes and their relations. Finally, the patent text vectors, structural feature vectors, and graph feature vectors are spliced together to predict target domain patents transfer or not. The performance of the model is verified by some comparative experiments. The experimental results show that the proposed model outperforms significantly the baseline models in the values of Precision, Recall, and F1 of patent data improve by 6.47%, 6.54%, and 6.58% on average, respectively.

A fully automated model for land use classification from historical maps using machine learning

Digital land use data before the age of satellites is scarce. Here, we build a machine learning model, using Extreme Gradient Boosting, that can automatically detect land use classes from an orthophoto map of Sweden (economic maps, 1:10 000 and 1:20 000) constructed from 1942 to 1988. Overall, the machine learning model demonstrated robust performance, with Cohen's Kappa and Matthews Correlation Coefficient of 0.86. The F1 values of the individual classes were 0.98, 0.95, 0.84, and 0.87 for graphics, arable land, forest, and open land, respectively. While the model can be used to detect land use changes in arable land, higher uncertainties associated with forest and open land necessitate further investigation at regional scales or exploration of improved mapping techniques. The code is publicly available to enable easy adaptation for classifying other historical maps.

Efficient Headline Generation with Hybrid Attention for Long Texts

Headline generation aims to condense key information from an article or a document into a concise one-sentence summary. The Transformer structure is in general effective for such tasks, yet it suffers from a dramatic increase in training time and GPU consumption as the input text length grows. To address this problem, a hybrid attention mechanism is proposed. Both local and global semantic information among words are modeled in a way that significantly improves training efficiency, especially for long text. Effectiveness is not sacrificed; in fact, fluency and semantic coherence of the generated headlines are enhanced. Experimental results on an open benchmark dataset show that, compared to the baseline model’s best performance, the proposed model obtains a 14.7%, 16.7%, 14.4% and 9.1% increase in the F1 values of the ROUGE-1, the ROUGE-2, the ROUGE-L and the ROUGE-WE metrics, respectively. The semantic coherence of the generated text is also improved, as shown by a 2.8% improvement in the BERTScore’s F1 value. These results show that the effectiveness of the proposed headline generation model with the hybrid attention mechanism is also improved. The hybrid attention mechanism could provide references for relevant text generation tasks.

Prediction of heart disease based on logistic regression

Heart disease is a major threat to human health, with a variety of contributing factors, and is not easily cured. This paper will present a dataset from a cardiovascular study of residents of Framingham, Massachusetts. First, the validity of the three models, logistic regression, random forest, and decision tree, is estimated by comparing information such as accuracy, precision, recall, and F1 values. The optimal model, i.e., the logistic regression model, was selected by plotting ROC curves and using AUC as a reference criterion for assessing the predictive effectiveness of the models. Then the raw data and data were preprocessed, including dealing with missing values. Finally, a logistic regression model was developed to analyze the influencing factors of heart disease. The purpose of this study was to use the results of the logistic model to help doctors and patients in heart disease treatment. The results show that the model has a good predictive effect.

A New Approach of Adjustment Factor 2023 (NAAF23) for Modified Slope Mass Rating (M-SMR)

The Modified Slope Mass Rating (M-SMR) system is a SMR-based geomechanical classification system utilized for rock slope characterization in the Crocker Formation. The M-SMR rating is derived from the sum of the basic Rock Mass Rating (RMRb) and an adjustment factor. However, it has been observed that the parallelism correction parameter, F1, within both the M-SMR and SMR systems, can sometimes be overestimated, especially for toppling failures when the discontinuity dip direction (αj) is less than the slope dip direction (αs). This study was conducted on six rock-cut slopes to not only evaluate the production of a convincing F1 value but also to introduce a simplified New Approach of Adjustment Factor 2023 (NAAF23) diagram for the M-SMR. This adjustment factor (F) includes four correction parameters (F1, F2, F3, and F4), similar to those used in SMR, but modifies the calculation approach for F1. The calculation now involves subtracting the higher value from the lower value among the discontinuity dip, slope dip, or intersection line orientations. The symbols A, B, C, and D represent the subtracted values, with A and B used when the discontinuity dip direction is higher than the slope dip direction and vice versa, and C and D used when the intersection line is higher than the slope dip direction and vice versa. For plane failures, A or B becomes the value, while for wedge failures, C and D are used. For toppling failures, the formula is 180 − A or B if A or B is less than 180, and A or B − 180 if A or B is greater than 180, eliminating the need for absolute symbols. A comparison F1 calculation using SMR is also conducted. The results show that F1 values become more convincing when using NAAF23.

Prediction of heart disease based on logistic regression

Heart disease is a major threat to human health, with a variety of contributing factors, and is not easily cured. This paper will present a dataset from a cardiovascular study of residents of Framingham, Massachusetts. First, the validity of the three models, logistic regression, random forest, and decision tree, is estimated by comparing information such as accuracy, precision, recall, and F1 values. The optimal model, i.e., the logistic regression model, was selected by plotting ROC curves and using AUC as a reference criterion for assessing the predictive effectiveness of the models. Then the raw data and data were preprocessed, including dealing with missing values. Finally, a logistic regression model was developed to analyze the influencing factors of heart disease. The purpose of this study was to use the results of the logistic model to help doctors and patients in heart disease treatment. The results show that the model has a good predictive effect.

Early diagnosis of Cladosporium fulvum in greenhouse tomato plants based on visible/near-infrared (VIS/NIR) and near-infrared (NIR) data fusion

Plant diseases can inflict varying degrees of damage on agricultural production. Therefore, identifying a rapid, non-destructive early diagnostic method is crucial for safeguarding plants. Cladosporium fulvum (C. fulvum) is one of the major diseases in tomato growth. This work presents a method of data fusion using two hyperspectral imaging systems of visible/near-infrared (VIS/NIR) and near-infrared (NIR) spectroscopy for the early diagnosis of C. fulvum in greenhouse tomatoes. First, hyperspectral images of samples at health and different times of infection were collected. The average spectral data of the image regions of interest were extracted and preprocessed for subsequent spectral datasets. Then different classification models were established for VIS/NIR and NIR data, optimized through various variable selection and data fusion methods. The principal component analysis-radial basis function neural network (PCA-RBF) model established using low-level data fusion achieved optimal results, achieving accuracies of 100% and 99.3% for calibration and prediction, respectively. Moreover, both the macro-averaged F1 (Macro-F1) values reached 1, and the geometric mean (G-mean) values reached 1 and 1, respectively. The results indicated that it was feasible to establish a PCA-RBF model by using the hyperspectral technique with low-level data fusion for the early detection of C. fulvum in greenhouse tomatoes.

Bunyi Forman Vokal dalam Bahasa Jawa Ngoko: Kajian Fonetik Akustik

The Javanese language (BJ) is divided into three, namely Ngoko, Krama Madya, and Krama Inggil. BJ consists of 10 vowel sounds: [i, I, e, ɛ, a, ǝ, o, ɔ, U, u]. This study aims to determine the values of BJ vowel sounds based on FORMAN data analysis, comparing the voices of male speakers (PL) and female speakers (PP). Additionally, the study aims to identify the more dominant vowel sounds based on gender. The qualitative method was used to collect data on the vowel sounds of Ngoko Javanese (BJN) by analyzing voice recordings using the Praat application program. The quantitative method was employed to calculate the FORMAN F1, F2, and F3 values and determine the average difference in BJN vowel sounds. Data from the BJ dictionary, consisting of 10 words with vowel sounds at the beginning, were used for analysis. The results showed that in PL, the highest vowel values were: [u] at 3089Hz, [i] at 3918Hz, [ɔ] at 3535Hz, and [ɛ] at 3228Hz. The lowest vowel values were: [ɛ] at 277Hz, [U] at 431Hz, and [i] at 336Hz. In PP, the highest vowel values were: [ɛ] at 3021Hz, [i] at 3200Hz, [I] at 3083Hz, and [o] at 3367Hz. The lowest vowel values were: [e] at 391Hz, [I] at 525Hz, [ɛ] at 385Hz, and [i] at 455Hz. In conclusion, BJN consists of 10 vowel sounds, with the highest vowel sound being [ɔ] and the lowest vowel sound being [i] in PL. The comparison of vowel sounds at the beginning of words in PL and PP in BJN, based on the high and low sounds of the F1, F2, and F3 vowel sounds, indicates that PL is more dominant.

Efficient degradation of haloacetic acids by vacuum ultraviolet-activated peroxymonosulfate: Kinetics, mechanisms and theoretical calculations

Efficient degradation of haloacetic acids (HAAs) is crucial due to their potential risks. This study firstly proposed vacuum ultraviolet - activated peroxymonosulfate (VUV/PMS) to remove HAAs (i.e., monochloroacetic acid (MCAA), monobromoacetic acid (MBAA), dichloroacetic acid (DCAA), etc). VUV/PMS achieved 99.51 % MCAA and 63.29 % TOC removal within 10 min. Electron paramagnetic resonance (EPR), quenching and probe experiments demonstrated that •OH was responsible for MCAA degradation. MCAA degradation followed pathways of dehalogenation (major) and decarboxylation (minor). VUV/PMS showed application potential under various reaction parameters. Broad spectrum of VUV/PMS on various HAAs was further explored. Chlorinated HAAs (Cl-HAAs) were primarily degraded by oxidation reactions, while brominated HAAs (Br-HAAs) by direct VUV photolysis. The density functional theory-based calculations (DFT) revealed that reaction rates of HAAs correlated with the highest occupied molecular orbital (HOMO) and energy gap (ΔE), indicating that HAAs degradation depends on their chemical structures. The Fukui function (f0 values) and bond length showed vulnerability of the halogen atom in Cl-HAAs and C–Br bond in Br-HAAs. Overall, this study provides an in-depth perspective on the oxidation performance and mechanism of HAAs using VUV/PMS. It not only demonstrates a green and efficient method but also inspires new strategies for HAAs remediation.

Balancing Acts: Tackling Data Imbalance in Machine Learning for Predicting Myocardial Infarction in Type 2 Diabetes.

Type 2 Diabetes (T2D) is a prevalent lifelong health condition. It is predicted that over 500 million adults will be diagnosed with T2D by 2040. T2D can develop at any age, and if it progresses, it may cause serious comorbidities. One of the most critical T2D-related comorbidities is Myocardial Infarction (MI), known as heart attack. MI is a life-threatening medical emergency, and it is important to predict it and intervene in a timely manner. The use of Machine Learning (ML) for clinical prediction is gaining pace, but the class imbalance in predictive models is a key challenge for establishing a trustworthy deployment of the technology. This may lead to bias and overfitting in the ML models, and it may cause misleading interpretations of the ML outputs. In our study, we showed how systematic use of Class Imbalance Handling (CIH) techniques may improve the performance of the ML models. We used the Connected Bradford dataset, consisting of over one million real-world health records. Three commonly used CIH techniques, Oversampling, Undersampling, and Class Weighting (CW) have been used for Naive Bayes (NB), Neural Network (NN), Random Forest (RF), Support Vector Machine (SVM), and Ensemble models. We report that CW overperforms among the other techniques with the highest Accuracy and F1 values of 0.9948 and 0.9556, respectively. Applying the most appropriate CIH techniques for the ML models using real-world healthcare data provides promising results for helping to reduce the risk of MI in patients with T2D.