Machine Learning

Iran is facing low levels of all three types of children's nutrition like nutrient and micronutrients deficiency and overeating. The most common nutritional problems and child deaths are vitamin deficiencies and food quality. The purpose of this research is to plan food recommended system to control malnutrition in children 6 to 12 years old using hybrid machine learning algorithms.  The results of this research are applicable in terms of target research. In terms of the implementation method, it is a descriptive survey and the process of gathering information is quantitative data. The dataset used includes 1001 data points collected from the health centers of Mianeh city located in East Azerbaijan in Iran from the integrated apple web system. In this research, the Python programming language has been used to analyze the child nutrition dataset, and AdaBoost and Decision Tree hybrid algorithms have been utilized for the child nutrients recommender system. We concluded that the number of meal features using the Decision Tree algorithm with 98.5% accuracy was more important than other nutritional features of children in recognizing malnutrition in them. From a review of 1001 data into the child nutrition dataset, 807 children are underweight and malnourished, 170 children are normal weight, 20 children are obese and four children are overweight. Therefore, the high exactness of hybrid algorithms in these studies has been able to have a high alignment with the opinion of nutritionists from 2019 to 2020.

Falls are a serious concern among the elderly people, causing severe physical pain to them and placing a strain on medical infrastructure. The global elderly population is expected to grow significantly in the coming years, as advances in healthcare allow lifespans to increase globally. This will bring more chances for falls to occur. With this in mind, there is a need for new research to be conducted on finding ways to reduce this problem. One area which shows promise is the use of Machine Learning to perform fall detection. Machine Learning is a rapidly growing field, and it has many applications in various fields such as finance, technology and medicine. When it comes to fall detection, Machine Learning systems are often able to detect falls much better and efficiently than a human can, given the same input data. The goal of this paper is to conduct a survey study on the main and most common machine learning algorithms implemented in the field of early fall detection for elderly people and the characteristics. The paper will discuss the different types of fall detection systems, algorithms, tools, datasets, applications, and challenges.  By conducting this research, a better understanding of the context, progress and trends in the field will be possible so that future research will have a guide to build upon.

“AppTree” is an intelligent platform to bring researchers, visitors, and all interested people closer to the oldest and most attractive botanical garden at the University of Tehran. AppTree can scan the QR-Barcode of each plant in person by smartphone or search various plants on the website and get all the useful knowledge about them. Also, the ability of AppTree is the recognition of different plants which don’t have labels. The plant recognition part is a machine learning module that can identify more than 100 different species of plants and give the user details about them. This novel platform is based on Android and Web-app and the identification of new plants type is done by machine learning approach. We utilized VGG19, a deep CNN, to classify images and to identify unlabeled plants. The classification accuracy, F1-score, recall, and precision were 98.25, 93.16%, 88.21%, and 94.85%, respectively, on the plant dataset of the University of Tehran. The proposed method was compared with other deep learning architectures such as AlexNet, AlexNetOWTBn, and GoogLeNet on the same dataset and obtained higher performance. Our AppTree platform has achieved considerable success and easily can be extended to use in other botanical gardens.

Community Question Answering has a crucial role in almost all societies nowadays. It is important for the owners of a community to be able to make it better and more reliable. One way to achieve this, is to find the users who have more knowledge, expertise, experience and skill and can well share their knowledge with others (which we call experts and aim to encourage them to be more active in the website). One method to use is to identify expert users, and whenever a new question is asked, we suggest this question to them to check and answer if its in their area of expertise. One way to encourage users to post replies, is to use gameplay techniques such as assigning points and badges to users. But as we will discuss, this method does not always detect expert users well, because some users will try to have small and insignificant but numerous activities that will make them gain a lot of points, however they are not experts. In this study, we examine the methods by which experts in a question-and-answer system can be found, and try to evaluate and compare these methods, use their ideas and positive points, and add our own new ideas to a new way of finding them. We used some ideas such as profile making for users, categorize users’ expertise, A-Priori algorithm and showed that neural networks method results the best for the purpose of expert detection.

Falls are a serious concern among the elderly due to being a major cause of harm to their physical and mental health. Despite their potential for harm, they can be prevented with proper care and monitoring. As such, the motivation for this research is to implement an algorithmic solution to the problem of falls that leverages the benefits of Machine Learning to detect falls in the elderly. There are various studies on fall detection that works on one dataset: wearable, environmental, or vision. Such an approach is biased against low fall detection and has a high false alarm rate. According to the literature, using two datasets can result in high accuracy and lower false alarms. The purpose of this study is to contribute to the field of Machine Learning and Fall Detection by investigating the optimal ways to apply common machine and deep learning algorithms trained on multimodal fall data. In addition, it has proposed a multimodal approach by training two separate classifiers using both Machine and Deep Learning and combining them into an overall system using sensor fusion in the form of a majority voting approach. Each trained model outputs an array comprising three percentage numbers, the average of the numbers in the same class from both arrays is then computed, and the highest percentage is the classification result. The working system achieved results were 97% accurate, with the highest being achieved by the Convolutional Neural Network algorithm. These results were higher than other state-of-the-art research conducted in the field.

The ability to predict the stock market and analyze market trends is invaluable to researchers and anyone interested in investing. However, this task is a challenging problem due to a large number of parameters and unpredictable noise that may affect the stock price. To overcome this issue, researchers have employed numerous approaches such as Moving Average (MA), Support Vector Machine (SVM), and Neural Networks. With technological advances, deep learning methods have become popular in processing time-series data. In this paper, we compare two recently introduced deep learning models, namely a Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), in forecasting daily movements of the Standard & Poor (S&P 500) index using the daily closing price of this index from 14/5/1991 to 14/5/2021. Results show that both models are effective and accurate in stock market prediction. In this case study, the mean squared error (MSE) and mean absolute error (MAE) for the GRU model are slightly lower than the LSTM model; hence, GRU outperformed the LSTM model despite its simpler structure. The results of this study are applicable in various instances where it is challenging to identify patterns among large volumes of unstructured data, such as medical data analysis, text mining, and financial time series modeling.

The issue of asset pricing in the market is one of the most important and old issues in the financial world. Factor pricing models seek to be able to determine a significant relationship between return on assets based on the risk parameters of that asset. A wide range of factors can be found in the literature that can be an element for measuring the risk of an asset, but the big question is which of these models will work better. The factors studied in this research include factors that cover market risk, valuation risk, psychological (technical) market risk, profit quality risk, profitability, investment, etc. In this study, we have tried to Using machine learning techniques and optimization tools is a way to derive adaptive-robust nonlinear models that can reduce the risk of model error as much as possible. In this research, two models have been developed. In the first model, using the feature extraction technique and optimization of models based on neural network, a non-linear and adaptable model has been developed for each asset. In the second approach, a portfolio of improved neural network-based models is used in the first stage, which can be used to minimize the risk of model error and achieve a model that is resistant to different market conditions. Finally, it can be seen that the development of these models can significantly improve the risk of error and average error of the model compared to traditional CAPM approaches and the Fama and French three-factor model.

Agriculture is one of the essential sources of occupation and revenue in India. Conferring to existing statistics, most agriculturalists are facing severe losses due to poor farming yield. Farming activities are challenged by various environmental factors that affect agricultural productivity to a greater extent. The present farming situation is above the average of the process involves more biochemical bases for managing the diseases and other destructing facts. The foremost problems they are facing in day-to-day farming tasks are crop or plant diseases affecting productivity. Also, the growth of weeds along with field crops has been another challenge.  The technology has developed to rectify the problems using some machine learning algorithms like Random Forest algorithms, Decision trees, Naïve Bayes, KNN, K-Means clustering, Support vector machines. The result has been evaluated and observed through the performance evaluation metrics using confusion matrix, accuracy, precision, Sensitivity, specificity with the observations, research, and studies. The statistics have expressed the overall accuracy of 98% by achieving the detection of diseases in plants and by removing the weeds that ruin the growth of plants.

Digital twin-enabled neural networks will develop innovative processes in feature selection and simulation. In addition, this methodology will have development in autonomous driving, natural language processing, healthcare, and many other fields. Recently sensors have been widely used for environment monitoring, and massive data has to be processed efficiently and effectively, which requires managed neural architectures for sustainable computing. The sustainable digital twin-empowered architectures create new biological evolution simulation algorithms and intelligent system architectures for supervised and unsupervised learning. Some of today's fundamental artificial intelligence issues, including adaptive machine learning and neuromorphic cognitive models, can be overcome by this methodology. The goals of this special issue on digital twin-enabled neural network architecture management for sustainable computing aim to pay attention to the researchers and industries towards recent advances in decision-making algorithms, neural network models and architectures for faster processing.

Extracting sentiments from the English-Telugu code-mixed data can be challenging and is still a relatively new research area. Data obtained from the Twitter API has to be in English-Telugu code-mixed language. That data is free-form text, noisy, lexicon borrowings, code-mixed, phonetic typing and misspelling data. The initial step is language identification and sentiment class labels assigned to each tweet in the dataset. The second step is the data normalization task, and the final step is classification, which can be achieved using three different methods: lexicon, machine learning, and deep learning. In the lexicon-based approach, tokenize each tweet with its language tag. If the language tag is in Telugu, transliterate the roman script into native Telugu words. Words are verified with TeluguSentiWordNet, and the Telugu sentiments are extracted, and English SentiWordNets are used to extract sentiments from the English tokens. In this paper, the aspect-based sentiment analysis approach is suggested and used with normalized data. In addition, deep learning and machine learning techniques are applied to extract sentiment ratings, and the results are compared to prior work.

In the past, deciding over granting loans to bank customers in Iran would be made traditionally and based on personal judgments over the risk of repayment. However, increase in demands on banking facilities by economic enterprises and families on the one side, and increased as well as extended commercial competitions among banks and financial and credit institutions in the country for reduction of facility repayment risk on the other side, have caused application of novel methods such as some statistical ones in this context. Now to predict the risk of negligence in banking facility repayment and classification of the candidates, bankers use their customers’ credit ranking. Time efficiency, cost effectiveness, avoidance from personal judgments, and further accuracy in examining the candidates who apply for various funds are of its salient merits of this new combined method. Various statistical methods including biased analysis, logistic regression, non-parametric parallelism, and also some others such as neural networks have been employed for credit ranking. In this research, given the random forest metaheuristic algorithm-based smart pattern of real bank customers’ credit risk (case study: Bank Tejarat) was presented. According to the value of skewness, the data could be stated to have a normal distribution. Based on the observed results, the lowest mean was related to the variable of type of facility and its maximum value, to the amount of facility.

What is clear is that judges usually judge cases based on their knowledge, experience, personality, and sentiment. Due to high pressures and stress, it may be difficult for them to carefully examine documents and evidence, which leads to more subjective judgments. Legal judgment prediction with artificial intelligence algorithms can benefit judicial bodies, legal experts, and litigants as well as judges. In this research, we are looking at predicting legal sentences in drug cases involving the purchase, possession, concealment, or transportation of illicit drugs, using machine learning methods, and the effect of sentiment and emotions in case texts on predicting the severity of whipping, fines, and imprisonment. So, the text documents of 6000 Persian drug-related cases were pre-processed and then the translation of the NRC Glossary of Emotions and sentiment was used to give each item a score for positive or negative sentiment and a score for emotion. Then machine learning methods were used for modeling. BERT, TFIDF+Adaboost, and Skipgram+LSTM+CNN methods had the highest accuracy, respectively. Also, evaluation criteria were analyzed in situations where sentiment scores, emotional scores, or both were used in the prediction process along with judicial texts. Finally, it was found that the use of sentiment and emotion scores improves the accuracy of legal judgment predictions for all three types of sentences and that sentiments have a greater impact on the accuracy of legal judgment predictions than emotions

Diabetes is a type of metabolic disorder with a high level of blood glucose. Due to the high blood sugar, the risk of heart-related diseases like heart attack and stroke got increased. The number of diabetic patients worldwide has increased significantly, and it is considered to be a major life-threatening disease worldwide. The diabetic disease cannot be cured but it can be controlled and managed by timely detection. Artificial Intelligence (AI) with Machine Learning (ML) empowers automatic early diabetes detection which is found to be much better than a manual method of diagnosis. At present, there are many research papers available on diabetes detection using ML techniques. This article aims to outline most of the literature related to ML techniques applied for diabetes prediction and summarize the related challenges. It also talks about the conclusions of the existing model and the benefits of the AI model. After a thorough screening method, 74 articles from the Scopus and Web of Science databases are selected for this study. This review article presents a clear outlook of diabetes detection which helps the researchers work in the area of automated diabetes prediction.

Internet-based social channels have turned into an important information repository for many people to get an idea about current trends and events happening around the world. As a result of Abundance of raw information on these social media platforms, it has become a crucial platform for businesses and individuals to make decisions based on social media analytics. The ever-expanding volume of online data available on the global network necessitates the use of specialized techniques and methods to effectively analyse and utilize this vast amount of information. This study's objective is to comprehend the textual information at the Lexical and Semantic level and to extract sentiments from this information in the most accurate way possible. To achieve this, the paper proposes to cluster semantically related words by evaluating their lexical similarity with respect to feature and sequence vectors. The proposed method utilizes Natural Language Processing, semantic and lexical clustering and hybrid C4.5 algorithm to extract six subcategories of emotions over three classes of sentiments based on word-based analysis of text. The proposed approach has yielded superior results with seven existing approaches in terms of parametric values, with an accuracy of 0.96, precision of 0.92, sensitivity of 0.94, and an f1-score of 0.92.

In healthcare, a diagnosis is reached after a thorough physical assessment and analysis of the patient's medicinal history, as well as the utilization of appropriate diagnostic tests and procedures. 1.7 million People worldwide lose their lives every year due to complications from chronic kidney disease (CKD). Despite the availability of other diagnostic approaches, this investigation relies on machine learning because of its superior accuracy. Patients with chronic kidney disease (CKD) who experience health complications like high blood pressure, anemia, mineral-bone disorder, poor nutrition, acid abnormalities, and neurological-complications may benefit from timely and exact recognition of the disease's levels so that they can begin treatment with the most effective medications as soon as possible. Several works have been investigated on the early recognition of CKD utilizing machine-learning (ML) strategies. The accuracy of stage anticipations was not their primary concern. Both binary and multiclass classification methods have been used for stage anticipation in this investigation. Random-Forest (RF), Support-Vector-Machine (SVM), and Decision-Tree (DT) are the prediction models employed. Feature-selection has been carried out through scrutiny of variation and recursive feature elimination utilizing cross-validation (CV). 10-flod CV was utilized to assess the models. Experiments showed that RF utilizing recursive feature removal with CV outperformed SVM and DT.

Heart failure is a severe medical ailment that significantly impacts patients’ well-being and the healthcare system. For improved results, early detection and immediate treatment are essential. This work aims to develop and evaluate predictive models by applying sophisticated ensemble learning techniques. In order to forecast heart failure, we used a clinical dataset from Kaggle. We used the well-known ensemble techniques of bagging and random forest (RF) to create our models. With a predicted accuracy of 82.74%, the RF technique, renowned for its versatility and capacity to handle complex data linkages, fared well. The bagging technique, which employs several models and bootstrapped samples, also demonstrated a noteworthy accuracy of 83.98%. The proposed model achieved an accuracy of 90.54%. These results emphasize the value of group learning in predicting cardiac failure. The area under the ROC curve (AUC) was another metric to assess the model’s discriminative ability, and our model achieved 94% AUC. This study dramatically improves the prognostic modeling for heart failure. The findings have extensive implications for clinical practice and healthcare systems and offer a valuable tool for early detection and intervention in cases of heart failure.

Heart and circulatory system diseases are often referred to as cardiovascular disease (CVD). The health and efficiency of the heart are crucial to human survival. CVD has become a primary cause of demise in recent years. According to data provided by the World-Health-Organization (WHO), CVD were conscientious for the deaths of 18.6M people in 2017. Biomedical care, healthcare, and disease prediction are just few of the fields making use of cutting-edge skills like machine learning (ML) and deep learning (DL). Utilizing the CVD dataset from the UCI Machine-Repository, this article aims to improve the accuracy of cardiac disease diagnosis. Improved precision and sensitivity in diagnosing heart disease by the use of an optimization algorithm is possible. Optimization is the process of evaluating a number of potential answers to a problem and selecting the best one. Support-Machine-Vector (SVM), K-Nearest-Neighbor (KNN), Naïve-Bayes (NB), Artificial-Neural-Network (ANN), Random-Forest (RF), and Gradient-Descent-Optimization (GDO) are just some of the ML strategies that have been utilized. Predicting Cardiovascular Disease with Intelligence, the best results may be obtained from the set of considered classification techniques, and this is where the GDO approach comes in. It has been evaluated and found to have an accuracy of 99.62 percent. The sensitivity and specificity were likewise measured at 99.65% and 98.54%, respectively. According to the findings, the proposed unique optimized algorithm has the potential to serve as a useful healthcare examination system for the timely prediction of CVD and for the study of such conditions.

Sentiment analysis is the process of identifying and categorizing people’s emotions or opinions regarding various topics. The analysis of Twitter sentiment has become an increasingly popular topic in recent years. In this paper, we present several machine learning and a deep learning model to analysis sentiment of Persian political tweets. Our analysis was conducted using Bag of Words and ParsBERT for word representation. We applied Gaussian Naive Bayes, Gradient Boosting, Logistic Regression, Decision Trees, Random Forests, as well as a combination of CNN and LSTM to classify the polarities of tweets. The results of this study indicate that deep learning with ParsBERT embedding performs better than machine learning. The CNN-LSTM model had the highest classification accuracy with 89 percent on the first dataset and 71 percent on the second dataset. Due to the complexity of Persian, it was a difficult task to achieve this level of efficiency. The main objective of our research was to reduce the training time while maintaining the model's performance. As a result, several adjustments were made to the model architecture and parameters. In addition to achieving the objective, the performance was slightly improved as well.