Research Projects
In light of my profound interest in applying my knowledge in the realms of data mining, machine learning, and deep learning to real-world scenarios, particularly within the field of healthcare and medicine, I showcase a selection of my research endeavors on this page. These endeavors aim to explore the practical applications of these methodologies and their implications in various contexts.
We are proud to have received funding from Abadan University of Medical Sciences, as it enabled us to carry out this project. The financial support provided by the University was crucial in facilitating the necessary resources, such as equipment, materials, and personnel, required for the successful completion of the project.
Using machine learning algorithms to predict successful aging: comparing basic vs ensemble techniques(2022)
This research project was commissioned by the Abadan University of Medical Sciences, Iran and tries to identify the best model for predicting successful aging and the quality of life of the elderly through the training of selected data mining models.
Aging is a chief risk factor for most chronic illnesses and infirmities. The growth in the aged population increases medical costs, thus imposing a heavy financial burden on families and communities. Successful aging (SA) is a positive and qualitative view of aging. From a biomedical perspective, SA is defined as the absence of diseases or disability disorders. This is distinct from normal aging, which is associated with age-related deterioration in physical and cognitive functions. From a social perspective, SA highlights life satisfaction and individual well-being, usually attained through socialization. It is an abstract and multidimensional concept surrounded by imprecision about its definition and measurement. This research project attempted to find the most effective features of SA as defined by Rowe and Kahn’s theory. The determined features were used as input parameters of six machine learning (ML) algorithms to create and validate predictive models for SA.
In this retrospective study, the raw data collected from the regional registry database of Khuzestan Universities of Medical Sciences, Iran was first pre-processed; then, based on the data of a sample of 983, five basic ML techniques including artificial neural network, decision tree, support vector machine, Naive Bayes, and k-nearest neighbors (K-NN) with one ensemble method (that gathers 30 K-NN algorithms as weak learners) were trained. The experimental results showed that ML techniques potentially have satisfactory performance in supporting the SA-related decisions of social and health policymakers. The KNN-based ensemble algorithm is superior to the other ML models in classifying people into SA and non-SA classes. Some of the results of our research projects have been published in the following papers:
- Which are best for successful aging prediction? Bagging, boosting, or simple machine learning algorithms?
- Prediction of successful aging using ensemble machine learning algorithms
Using Adaptive Neuro-Fuzzy Inference System for Predicting Successful Aging (2022)
This research project was commissioned by the Abadan University of Medical Sciences, Iran.
Considering the change in the age trend of people in society and the aging of the population, creating support programs for the elderly in the form of predictive models and expert classifiers through machine learning techniques is effective in improving the quality of life and quality indicators of people. Based on the indicators of the World Health Organization, people over 60 years old are considered elderly. The concept of successful aging (SA) includes three dimensions physiological, psychological, and social functioning. It has been proven that the use of machine learning methods to predict the medical-social conditions of the elderly has been effective in improving their quality of life indicators. Therefore, this research has tried to identify the best structure for predicting successful aging and the quality of life of the elderly retrospectively and through the training of selected data mining models. This research was conducted with the aim of evaluating the performance of a method based on fuzzy logic (ANFIS) for predicting SA. The data collected from the registry database of regional health centers of Abadan University of Medical Sciences were pre-processed before being used for modeling, outliers were identified and missing values were addressed. After analyzing the role of input variables in predicting the outcome, the ANFIS model was implemented to predict the quality of life of old age and successful old age. The prediction results showed that this model can provide a reliable and responsive tool for improving the outcomes of the elderly to geriatricians, geriatric nurses, health care administrators, and policymakers.
Deep text clustering using stacked AutoEncoder (2022)
This research project was commissioned by Iran's Barez Rubber Company. Textual data is a form of unstructured information that is easily processed by humans but poses challenges for computers to comprehend. Text mining techniques effectively discover meaningful insights from text, garnering significant attention in recent years. The purpose of this research was to evaluate and analyze opinions and suggestions provided by the Barez Iranian company. The dataset from the company was unlabeled. Manual extraction of useful information from large unlabeled textual data is very difficult and time-consuming. Therefore, this study employed intelligent deep-learning methods to analyze the suggestions presented in the Persian language. Using the BERTopic model, the dataset was analyzed through clustering, assigning each document to a topic with a probability distribution. Additionally, a novel deep text clustering approach, combining a stacked autoencoder encoder and k-means clustering, was proposed to organize the textual documents into semantically meaningful groups for information extraction from the prominent dataset. Experimental evaluations demonstrated that the proposed model notably outperforms other clustering methods. Some of the results of our research projects have been published in the following paper:
Deep text clustering using stacked AutoEncoder
Extracting knowledge from the data of diabetic patients using Metaheuristic algorithms and Data mining models (2018)
This research project was commissioned by Kosar University of Bojnord,Iran. Extracting knowledge from the massive amount of data related to disease records and patients' medical records using the data mining process can lead to the identification of the rules governing the development of diseases. It provides valuable information to healthcare professionals for identifying disease causes, predicting outcomes, and undertaking appropriate treatments. Diabetes is the fourth leading cause of death in most developed countries. The main objective of this research is to present an algorithm for knowledge extraction and pattern discovery from diabetic patient data to achieve more accurate and efficient early prediction of diabetes. To this end, clustering algorithms were utilized to eliminate noisy and outlier data. Additionally, an improved metaheuristic algorithm was used for selecting influential features in disease diagnosis. The performance of the proposed algorithm was evaluated on a diabetes dataset using a perceptron neural network. The results demonstrated the superiority of the proposed model over other algorithms in this research project.