Background and Purpose: Nitrates have long been considered indicative of drinking water quality and a critical concern for human health. The evolution of advanced models for water quality management has spurred decision-makers to incorporate artificial intelligence technologies into water quality planning. This study aims to employ the AdaBoost model, one of the cutting-edge models in water quality management, to predict nitrate concentrations in groundwater using pH and EC (Electrical Conductivity) as input variables. Materials and Methods: Initially, the study analyzed the Pearson correlation matrix and subsequently determined the input variables for multiple AdaBoost models with varying hyperparameters. A sensitivity and dependence analysis of the model's input variables was conducted to assess their impact on nitrate prediction. Results: The results obtained from the AdaBoost model reveal R-squared (R2) values of 0. 915 for the training dataset and 0. 924 for the test dataset. Additionally, the Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) scores for the training dataset were recorded as 1. 02, 1. 01, 0. 823, and 7. 3%, respectively. For the test dataset, these metrics were observed in the order of 0. 228, 0. 477, 0. 375, and 3. 2%. The model's sensitivity analysis identified the pH variable as the most influential factor in nitrate prediction. Conclusion: The model analysis demonstrates that the proposed method performs well in predicting nitrate concentrations. This approach holds significant potential for implementation as an intelligent system for forecasting water quality parameters.

Introduction: Liver disease is one of the most common and dangerous diseases the early detection of which can be very effective in preventing complications as well as controlling and treating the disease. The purpose of this study was to improve AdaBoost algorithm using Firefly Algorithm for diagnosing liver disease. Method: This is a descriptive-analytic study. The dataset consists of 583 independent records including 10 features of machine learning dataset in the University of California, Irvine. In this study, AdaBoost and Firefly Algorithm were combined to increase the effectiveness of liver disease diagnosis. 80% of the data were used for training and 20% for testing. Results: The results highlighted the superiority of the hybrid model of feature selection over the models without feature selection. Of course, the selection of important features affect the performance of the model. The accuracy of the hybrid model considering 5 and all features was 98. 61% and 94. 15%, respectively. Overall, the hybrid model proved more accurate compared with most of the other data mining models. Conclusion: Hybrid model can be used to help physicians identify and classify healthy and unhealthy individuals; it can also be used in medical centers to enhance accuracy and speed, and reduce costs. It cannot be claimed that the hybrid model is the best model; however, it proved more accurarate.

Background: Temperament (Mizaj) determination is an important stage of diagnosis in Persian Medicine. This study aimed to evaluate thermal imaging as a reliable tool that can be used instead of subjective assessments. Methods: The temperament of 34 participants was assessed by a PM specialist using standardized Mojahedi Mizaj Questionnaire (MMQ) and thermal images of the wrist in the supine position, the back of the hand, and their whole face under supervision of the physician were recorded. Thirteen thermal features were extracted and a classifying algorithm was designed based on the genetic algorithm and AdaBoost classifier in reference to the temperament questionnaire. Results: The results showed that the mean temperature and temperature variations in the thermal images were relatively consistent with the results of MMQ. Among the three body regions, the results related to the image from Malmas were most consistent with MMQ. By selecting six of the 13 features that had the most impact on the classification, the accuracy of 94. 7 ±,13. 0, sensitivity of 95. 7 ±,11. 3, and specificity of 98. 2 ±,4. 2 were obtained. Conclusions: The thermal imaging was relatively consistent with standardized MMQ and can be used as a reliable tool for evaluating warm/cold temperament. However, the results reveal that thermal imaging features may not be only main features for temperament classification and for more reliable classification, it needs to add some different features such as wrist pulse features and some subjective characteristics.

The application of communication between human and computer has been emerging as an important matter in communication with the surrounding environment. If the computer could sense the human’ s emotion, it would be easier to establish a connection between the computer and human. Therefore, the extraction of emotions is an important topic in communication between them. For extracting an emotion, which is absolutely undeniable, various biological signals are used. One of the simple and high-precision methods to acquire data from these signals is to implement the eye tracking and the concentration on the screen technique. In this paper, the eye tracking technique is used to extract emotions for the communication between the human and computer. According to the acquired data from the persons and videos, some of the characteristics of signals, including focus areas, pupil diameter, statistical features, and features of videos are extracted. In addition, in order to improve the results, combining the features are proposed. Afterwards, based on two distinct outputs i. e. Arousal and Valence, and employing a linear combination and reducing the dimension, some features are selected separately. Finally, to classify the two-axes associated with the Arousal and Valence in the range of 0 to 9, which is divided into three equal parts; special types of KNN and SVM methods combined with AdaBoost classifier are used. The numerical studies have shown that the average extraction accuracy is 68. 66% for Arousal axis, and 74. 66% for Valence axis. As a result, the overall accuracy is improved 5. 5% compared to the previous works, respectively.

With the increasing spread of attacks on computer networks, the use of intrusion detection systems is inevitable. The purpose of an intrusion detection system is to monitor abnormal activities and to distinguish between normal and abnormal behaviors (intrusion) in a host system or in a network. One of the main problems of intrusion detection systems is the high volume of alarms, which practically eliminates the possibility of dealing with them. An intrusion detection system is effective that can detect a wide range of attacks while reducing the amount of false alarms. In this paper, a new feature-based intrusion detection approach is proposed in which the AdaBoost algorithm combines with the Harris Hawks optimization algorithm and optimized parameters. Studies show that the proposed method detects malicious samples in computer networks with an average accuracy in the CICIDS2017 dataset is 99.86% and for the NSL_KDD dataset is 99.88%; comparing the findings with similar works also indicates that the proposed method is more accurate than them in distinguishing invasive samples from normal.

The diagnosis of hypertensive retinopathy (CAD-RH) can be made by observing the tortuosity of the retinal vessels. Tortuosity is a feature that is able to show the characteristics of normal or abnormal blood vessels. This study aims to analyze the performance of the CAD-RH system based on feature extraction tortuosity of retinal blood vessels. This study uses a segmentation method based on clustering self-organizing maps (SOM) combined with feature extraction, feature selection, and the ensemble Adaptive Boosting (AdaBoost) classification algorithm. Feature extraction was performed using fractal analysis with the box-counting method, lacunarity with the gliding box method, and invariant moment. Feature selection is done by using the information gain method, to rank all the features that are produced, furthermore, it is selected by referring to the gain value. The best system performance is generated in the number of clusters 2 with fractal dimension, lacunarity with box size 22-29, and invariant moment M1 and M3. Performance in these conditions is able to provide 84% sensitivity, 88% specificity, 7. 0 likelihood ratio positive (LR+), and 86% area under the curve (AUC). This model is also better than a number of ensemble algorithms, such as bagging and random forest. Referring to these results, it can be concluded that the use of this model can be an alternative to CAD-RH, where the resulting performance is in a good category.

Counting mitotic figures present in tissue samples from a patient with cancer, plays a crucial role in assessing the patient’s survival chances. In clinical practice, mitotic cells are counted manually by pathologists in order to grade the proliferative activity of breast tumors. However, detecting mitoses under a microscope is a labourious, time-consuming task which can benefit from computer aided diagnosis. In this research we aim to detect mitotic cells present in breast cancer tissue, using only texture and pattern features. To classify cells into mitotic and non-mitotic classes, we use an AdaBoost classifier, an ensemble learning method which uses other (weak) classifiers to construct a strong classifier.11 different classifiers were used separately as base learners, and their classification performance was recorded. The proposed ensemble classifier is tested on the standard MITOS-ATYPIA-14 dataset, where a 64×64 pixel window around each cells center was extracted to be used as training data. It was observed that an AdaBoost that used Logistic Regression as its base learner achieved a F1 Score of 0.85 using only texture features as input which shows a significant performance improvement over status quo. It is also observed that "Decision Trees" provides the best recall among base classifiers and "Random Forest" has the best Precision.