Gestational diabetes is a type of high blood sugar that develops during pregnancy. It can occur at any stage of pregnancy and cause problems for both the mother and the baby, during and after birth. The risks can be reduced if they are early detected and managed, especially in areas where only periodic tests of pregnant women are available. Intelligent systems designed by machine learning algorithms are remodelling all fields of our lives, including the healthcare system. This study proposes a combined prediction model to diagnose gestational diabetes. The dataset was obtained from the Kurdistan region laboratories, which collected information from pregnant women with and without diabetes. The suggested model uses the clustering KMeans technique for data reduction and the elbow method to find the optimal k value and the Mahalanobis distance method to find more related cluster to new samples, and the classification methods such as decision tree, random forest, SVM, KNN, logistic regression, and Naïve Bayes are used for prediction. The results showed that using a mix of KMeans clustering, elbow method, Mahalanobis distance, and ensemble technique significantly improves prediction accuracy.
11 September 2022 in Journal of Studies in Science and Engineering
Diabetes mellitus is also called gestational diabetes when a woman has high blood sugar while pregnant. It can show up at any time during pregnancy and cause problems for the mother and baby during or after the pregnancy. If the risks are found and dealt with as soon as possible, there is a chance that they can be reduced. The healthcare system is one of the many parts of our daily lives that are being rethought thanks to the creation of intelligent systems by machine learning algorithms. In this article, a hybrid prediction model is suggested to determine if a woman has gestational diabetes. The recommended model reduces the amount of data using the K-means clustering method. Predictions are made using several classification methods, such as decision trees, random forests, SVM, KNN, logistic regression, and naive Bayes. The results show that accuracy increases when clustering and classification are used together.
Nowadays, due to the high volume of choices for customers which causes confusion, the use of recommender systems is strongly growing. Of course, existing systems have two problems, one is complexity and the other is failure to consider uncertainty. In this article, we have reduced the complexity of the system by using a fuzzy innovative system and solved the problem of the uncertainty of users' ratings regarding goods. For that purpose, this research attempts to extract fuzzy membership functions from the Yahoo movie dataset for recommendation applications. In the proposed method, a type I fuzzy system with low numbers of membership functions is designed. The uncertainty in users' ratings is handled by clustering users and movies. Moreover, repeated user evaluations of the same movies are used to determine the uncertainty in improved type 1 membership functions. To evaluate the proposed strategy, MAE, confusion matrix, and Classification-report are used. The result demonstrates the superiority of the introduced strategy.
A Recommender System (RS) is one of the appropriate answers by researchers for alleviating the problem of customers' overload with information from an internet source. In RS, users' evaluation of the items assists in quickly determining the user's preferences and removing the choosing overhead for the user the next time they search. However, using user ratings for an item based on multiple aspects has given researchers little attention to recommending an object accurately. This research proposes a model-based collaborative filtering movie recommendation system based on user ratings across various criteria. The proposed systems' model was established using clustering and classification methods, such as k-means, k-modes, and Multinomial logistic regression. The proposed work consists of two different steps; first, clustering mode; after clustering the dataset with k-means, k-modes were used to re-cluster it into many sub-clusters for search domain reduction purposes. The second is classification mode; the Multinomial Logistic Regression (MLR) model was created to predict the closest cluster class to newly active users. Since the MLR is one of the probabilistic models, it is more accurate than cluster methods in the prediction process. The proposed approach uses distance indicators, such as modified Mahalanobis distance and Euclidean distance, to measure the similarity between new active and in-group users. The evaluation of the proposed methodology was done using the MAE and Silhouette scores. Different values of Silhouette score were achieved in this study for a different number of features and clusters, with the best k-means score being 0.822. Based on multicriteria Yahoo!! Movie Dataset, the MAE result indicates that this study is better than the existing one.
General type two fuzzy models are rarely used in practical problems due to their computational complexity, especially during fuzzification and defuzzification. This is why the main purpose of this article is to present a simple and effective method with a closed formula to increase their use in real-time tasks. For this purpose, we interpreted the general type two membership function as a set of infinite horizontal layers of uncertainty footprints, and then a closed formula for defuzzification is calculated using the uncertainty avoidance method for the triangular shape of secondary membership degree forms. One of the abilities that is expected from Fuzzy logic, particularly General Type-2 Fuzzy Systems (GT2FS), is their effective role in noise handling. In this article, we emphasize providing a method that can be used more consciously to handle noise. To demonstrate the effectiveness of the proposed multilayer approach, it has been applied to image segmentation, a critical aspect of image processing often hindered by challenges like noise, low contrast, and blurred features. The proposed multilayer type-2 fuzzy method based on the uncertainty avoider method has shown very good performance in image segmentation, especially in the case of high-noise images, compared to type one Fuzzy C-Means (FCM) and type-2 FCM methods as well as hybrid methods presented recently in the literature.
This paper proposes a robust convolutional neural network (CNN) architecture for human activity recognition (HAR) using smartphone accelerometer data, evaluated on the WISDM dataset. We introduce two novel pooling mechanisms—Pooling A (Extrema Contrast Pooling (ECP)) and Pooling B (Center Minus Variation (CMV))—that enhance feature discrimination and noise robustness. ECP emphasizes sharp signal transitions through a nonlinear penalty based on the squared range between extrema, while CMV Pooling penalizes local variability by subtracting the standard deviation, improving resilience to noise. Input data are normalized to the [0, 1] range to ensure bounded and interpretable pooled outputs. The proposed framework is evaluated in two separate configurations: (1) a 1D CNN applied to raw tri-axial sensor streams with the proposed pooling layers, and (2) a histogram-based image encoding pipeline that transforms segment-level sensor redundancy into RGB representations for a 2D CNN with fully connected layers. Ablation studies show that histogram encoding provides the largest improvement, while the combination of ECP and CMV further enhances classification performance. Across six activity classes, the 2D CNN system achieves up to 96.84% weighted classification accuracy, outperforming baseline models and traditional average pooling. Under Gaussian, salt-and-pepper, and mixed noise conditions, the proposed pooling layers consistently reduce performance degradation, demonstrating improved stability in real-world sensing environments. These results highlight the benefits of redundancy-aware pooling and histogram-based representations for accurate and robust mobile HAR systems.