Showing papers by "Chittagong University of Engineering & Technology published in 2018"

Studies of ionizing radiation shielding effectiveness of silica-based commercial glasses used in Bangladeshi dwellings

Abstract: Following the rapid growing economy, the Bangladeshi dwellers are replacing their traditional (mud-, bamboo-, and wood-based) houses to modern multistoried buildings, where different types of glasses are being used as decorative as well as structural materials due to their various advantageous properties. In this study, we inquire the protective and dosimetric capability of commercial glasses for ionizing radiation. Four branded glass samples (PHP-Bangladesh, Osmania-Bangladesh, Nasir-Bangladesh, and Rider-China) of same thickness and color but different elemental weight fractions were analyzed for shielding and dosimetric properties. The chemical composition of the studied material was evaluated by EDX technique. A well-shielded HPGe γ-ray spectrometer combined with associated electronics was used to evaluate the attenuation coefficients of the studied materials for 59 keV, 661 keV, 1173 keV and 1332 keV photon energies. A number of shielding parameters- half value layer (HVL), radiation protection efficiency (RPE) and effective atomic number (Zeff) were also evaluated. The data were compared with the available literature (where applicable) to understand its shielding capability relative to the standard materials such as lead. Among the studied brands, Rider (China) shows relatively better indices to be used as ionizing radiation shielding material. The obtained, Zeff of the studied glass samples showed comparable values to the TLD-200 dosimeter, thus considered suitable for environmental radiation monitoring purposes.

Sensor Applications and Physiological Features in Drivers’ Drowsiness Detection: A Review

Abstract: Drowsiness in drivers has become a serious cause of concern due to the occurrences of a large number of fatalities on the road each year. Lives of pedestrians and passengers are put to risk as drivers tend to fall asleep at the steering wheel. In the recent past, many researchers have paid attention to the problem of drowsiness detection since safe roads and safe driving are of paramount concern to all societies. This paper has led to the development of several novel and effective methods in detecting drivers’ drowsiness. These include: 1) Vehicle based methods; 2) Behavioral methods; and 3) Physiological methods. Since wake-sleep is an intermediate state between two physiologically dissimilar states, physiological signals can define this transition more accurately when compared with approaches that fall in other categories. This paper focuses on the role of physiological signals in detecting driver’s drowsiness level. The proposed methods measure the physiological signals by means of various sensors, which monitor the driver’s physiological parameters on a continual basis. Multiple sensors can be embedded on the driver or in the vicinity of the driver to capture vital signs indicating the onset of drowsiness. The aim here is to provide an insightful review of all such key approaches that fall in this category. This paper conducts a detailed study in which key physiological parameters that relate to drowsiness are identified, described, and analyzed. Furthermore, the overall advantages and limitations of these physiological based schemes are also highlighted.

Improved photovoltaic performance from inorganic perovskite oxide thin films with mixed crystal phases

Abstract: Inorganic ferroelectric perovskites are attracting attention for the realization of highly stable photovoltaic cells with large open-circuit voltages. However, the power conversion efficiencies of devices have been limited so far. Here, we report a power conversion efficiency of ~4.20% under 1 sun illumination from Bi–Mn–O composite thin films with mixed BiMnO3 and BiMn2O5 crystal phases. We show that the photocurrent density and photovoltage mainly develop across grain boundaries and interfaces rather than within the grains. We also experimentally demonstrate that the open-circuit voltage and short-circuit photocurrent measured in the films are tunable by varying the electrical resistance of the device, which in turn is controlled by externally applying voltage pulses. The exploitation of multifunctional properties of composite oxides provides an alternative route towards achieving highly stable, high-efficiency photovoltaic solar energy conversion. The use of mixed crystal phases is shown to enhance the power conversion efficiency of inorganic ferroelectric perovskite solar cells.

The radiation shielding offered by the commercial glass installed in Bangladeshi dwellings

Abstract: One sign of a vibrant Bangladeshi economy has been the move away from the use of more traditional housing materials towards a preference for modern constructional media. Glass, one such example, used both decoratively and in a structural context, offers various advantageous properties and facets including a protective feature against radiation that has not previously been considered. Current interest examines the dosimetric possibilities offered by the commercial glass as a secondary shield and also in retrospective ionising radiation exposure analysis. Four popular brands of window glass are investigated, all available within the local market (PHP-Bangladesh, Usmania-Bangladesh, Nasir-Bangladesh and Rider-China), all with the same thickness and colour, varying in terms of elemental weight fractions as evaluated by energy dispersive X-ray analysis. As potential attenuators of transmitted radiation thereby forming secondary barriers against radiation exposure from penetrating radiations, the four b...

First principles study of M2InC (M = Zr, Hf and Ta) MAX phases: The effect of M atomic species

Abstract: We have studied the physical properties of M2InC (M = Zr, Hf and Ta) MAX phases ternary carbides using density functional theory (DFT) methodology. The structural, elastic and electronic properties are revisited (and found to be in good agreement with recently reported results). The charge density distribution, Fermi surface features, Vickers hardness, dynamical stability, thermodynamics and optical properties have been investigated for the first time. The calculated single crystal elastic constants and phonon dispersion curves endorse the mechanical and dynamical stability of all the compounds under study. The calculated single crystal elastic constants Cij and polycrystalline elastic constants are found to increase with increasing atomic number of M species (M = Zr, Hf and Ta). The values of Pugh ratio and Poisson’s ratio revealed the brittleness of the compounds under study associated with strong directional covalent bond with a mixture of ionic contribution. Overlapping of conduction band and valence band at Fermi level notify the metallic nature of M2InC (M = Zr, Hf and Ta) MAX phases. Low values of Vicker hardness indicate the softness of the materials and easy machinability. The thermodynamic properties, such as the free energy, enthalpy, entropy, specific heat capacity and Debye temperature are evaluated using the phonon dispersion curves and a good correspondence is found with the characteristics of M atomic species. Major optical properties, e.g., dielectric functions, refractive index, photoconductivity, absorption coefficient, loss function and reflectivity are calculated and discussed in detail in this study.

The presence of radioactive materials in soil, sand and sediment samples of Potenga sea beach area, Chittagong, Bangladesh: Geological characteristics and environmental implication

Abstract: Accurate quantification of naturally occurring radioactive materials in soil provides information on geological characteristics, possibility of petroleum and mineral exploration, radiation hazards to the dwelling populace etc. Of practical significance, the earth surface media (soil, sand and sediment) collected from the densely populated coastal area of Chittagong city, Bangladesh were analysed using a high purity germanium γ-ray spectrometer with low background radiation environment. The mean activities of 226Ra (238U), 232Th and 40K in the studied materials show higher values than the respective world average of 33, 36 and 474 Bq/kg reported by the UNSCEAR (2000). The deduced mass concentrations of the primordial radionuclides 238U, 232Th and 40K in the investigated samples are corresponding to the granite rocks, crustal minerals and typical rocks respectively. The estimated mean value of 232Th/238U for soil (3.98) and sediment (3.94) are in-line with the continental crustal average concentration of 3.82 for typical rock range reported by the National Council on Radiation Protection and Measurements (NCRP). But the tonalites and more silicic rocks elevate the mean value of 232Th/238U for sand samples amounting to 4.69. This indicates a significant fractionation during weathering or associated with the metasomatic activity in the investigated area of sand collection.

Design of an EEG-Based Brain Controlled Wheelchair for Quadriplegic Patients

Abstract: In this paper, an EEG-based brain controlled wheelchair has been designed using Brain Computer Interface (BCI) with the help of NeuroSky Mind Wave EEG Headset. Quadriplegic patients can not move any organ of their body below their necks. Implementation of this device will assist the quadriplegic patients to move independently on their own. The movement of the wheelchair is controlled by variation of attention level of the patient. Turning on or off this device is controlled by double eye blink of the patient. Graphics based fuzzy image has been added in the design of the wheelchair for assisting the patients varying their attention levels if needed.

Nonlinear time fractional Korteweg-de Vries equations for the interaction of wave phenomena in fluid-filled elastic tubes

Abstract: This work investigates the wave-wave interactions by considering two-sided time fractional Korteweg-de Vries equations (TFKdVEs). The generalized $\exp(-\Phi(\xi))$ -expansion method (GEEM) along with Khalil's fractional derivatives is employed to divulge several types of scattered wave solutions of TFKdVEs. It is found that the fractional parameters significantly modified the interaction of nonlinear wave dynamics. The results obtained may be useful for clarifications of the interaction between two waves not only in non-conservative fluid-filled elastic tubes but also in shallow water and plasma physics.

Feature Extraction from Dermoscopy Images for an Effective Diagnosis of Melanoma Skin Cancer

Abstract: The aim of this paper is to extract some distinct geometric features from dermoscopy images to classify benign and malignant melanomas. To avoid skin biopsy which is an invasive technique, diagnosis of melanoma skin cancer from dermoscopy images was developed. It is a very challenging task due to some reasons. Firstly, high degree of intraclass variation exists among melanoma images while low interclass variation is found between melanoma and non-melanoma images. Secondly, benign and malignant melanoma images are visually similar to a great extent. And finally noises like hair are always present in skin images which make difficult to analyse the images. In this work, we used fundamental ABCD rule to detect malignant melanoma and benign lesion based on quantitative measures. In our proposed technique, we extracted a new feature which is the difference between maximum and minimum Feret diameters of the best fit ellipse to skin lesion. This discriminative feature alone classified the melanomas with 86.5% accuracy. In our approach, we applied the feature extraction block containing all parameters to 200 images and the overall accuracy of 98% was achieved to detect malignant and benign melanoma from the images. A Back-propagation Neural Network (BNN) model was developed and eventually used as a classifier in this proposed method.

Automatic Sickle Cell Anemia Detection Using Image Processing Technique

Abstract: One of the important parts of the human body is red blood cells (RBCs). Disk shape is the ordinary red blood cell’s shape. One type of ailment of blood is sickle cell anemia (SCA) in where red blood cells are formed in crescent shapes from their actual shapes. Thousands of babies around the world are born with this blood disorder every year. The numbers of SCA are assumed to increase about 30% by 2050 globally. About 0.07 to 0.1 million Americans are victims of SCA. A new way is introduced to detect and classify sickle cells in RBC using image processing technique to start treatment as early as possible. Firstly, this method collects images of blood. The pre-processing phase is done through gray scale image conversion, image enhancement & median filter. Then, the threshold segmentation is applied to segment the RBCs and morphological operations are used to remove the undesired objects from images. Metric value, aspect ratio, entropy, mean, standard deviation and variance are used as features which are extracted. Finally, the support vector machine classifier is trained to test the images. The system provided better accuracy and sensitivity than existing methods. This automatic detection technique would be very useful to save the precious lives of the people.

Head-on collision of ion acoustic shock waves in electron-positron-ion nonextensive plasmas for weakly and highly relativistic regimes

Abstract: Head-on collision between ion acoustic shock waves (IASWs) and the consequences after collision are investigated considering the plasma system to be consisting of relativistic warm ions and nonextensive electrons and positrons, taking into account the effects of nonlinearity and dispersion. Two-sided KdV-Burger equations are derived employing the extended Poincar e ´-Lighthill-Kuo method. The results reveal that the plasma parameters are responsible for the modification of the structures along with phase shifts of the shock waves. The nonlinearity effects on IASWs in a highly relativistic regime (HRR) become pronounced rather than the weakly relativistic regime (WRR). The phase shifts of IASWs are enhanced by the relativistic streaming factor and superthermality. The electrostatic IASWs become rarefactive depending on temperatures, kinematic viscosity, and superthermality in both WRR and HRR. The amplitudes of IASWs are increasing for WRR but decreasing for HRR due to increasing ion thermal velocities. Besides, the amplitudes of the solitons are detaining due to the increase in the positron concentration for the depopulation of ions.

Utilization of E-waste in Concrete and its Environmental Impact - A Review

Abstract: Electronic waste or e-waste defines rejected electrical or electronic devices. Secondhand electronics which are destined for recover, resale, salvage, recycling, or disposal are also considered e-waste. Electronic waste is an emerging concern posing serious contamination problems to the human and the environment. E-waste disposal is a typical task for whole over the world. Utilization of E-waste materials is a partial solution to environmental and ecological problems. Due to large amount of concrete use as the construction material availability of raw material is being questioned. Therefore, other replacing materials are needed to be finding out. E-waste is used as one such alternative for aggregate in concrete. This paper presents a summary on probable use of E-waste in concrete on the basis of different researchers and it marches strong possibility of E-waste being used as additional of aggregate as well as environmental impact. The use of natural aggregates in concrete will be decreased if different types of by-product is used in concrete as a substitute material. And it is more important to renovate the waste material.

Yttrium-substituted Mg-Zn ferrites: correlation of physical properties with Yttrium content

Abstract: Yttrium (Y) substituted Mg Zn ferrites with the compositions of Mg0.5Zn0.5YxFe2 xO4 have been synthesized by conventional standard ceramic technique. The effect of Y3+ substitution on the structural, electrical, dielectric and magnetic properties of Mg Zn ferrites has been studied.

Smoke Detection Method Based on LBP and SVM from Surveillance Camera

Abstract: Wildfire is a regular incident worldwide today. It destroys forests and also the living areas of wild animals. So, to reduce the harmful effects of such disasters this paper describes a method of smoke detection for surveillance cameras. This smoke detection will ease the fire detection. Proposed method is based on Local Binary Pattern (LBP) and Support Vector Machine (SVM). Initially, Approximate Median Filtering Algorithm was applied to subtract the background from input frame. Then, shape based filtering method was applied to get the region of interest. Thirdly, LBP values and histograms were calculated from the pixels of region of interest to form a feature vector. The proposed method also applied Bhattacharyya coefficients to verify the smoke region for accurate result. Finally, SVM classified the region of interest as smoke image. Results using real scene data show that the proposed method can give accurate results in different conditions of real world situations.

Novel Visible-Light-Driven Photocatalyst Co 3 O 4 /FeWO 4 for Efficient Decomposition of Organic Pollutants

Abstract: A highly efficient and visible light (λ ≥ 420 nm) responsive composite photocatalyst, Co3O4/FeWO4 was prepared by simple impregnation method. The heterojunction semiconductors Co3O4/FeWO4 demonstrated notably high photocatalytic activity over a wide range of composition than the individual component Co3O4 or FeWO4 for the complete degradation of 1,4-dichlorobenzene (DCB) in aqueous phase under visible light irradiation. The photocatalytic activity of composite was optimized at 1/99 Co3O4/FeWO4 composition. After 2 h of visible light irradiation 51% decomposition of 1,4-dichlorobenzene (DCB) was observed utilizing 1/99 Co3O4/FeWO4 photocatalyst while the end members demonstrated a negligible degradation under the same experimental condition. The valence band (VB) and conduction band (CB) of Co3O4 is located above the VB and CB of FeWO4, respectively. Both the semiconductors Co3O4 and FeWO4 exhibit strong absorption over the wide range of visible light. The obviously enhanced photocatalytic performance of Co3O4/FeWO4 composite has been discussed on the hole (h+) as well as electron (e−) transfer mechanism between the VB and CB of individual semiconductors.

Effect of fly ash on freeze–thaw durability of concrete in marine environment

Abstract: Deterioration of structural concrete exposed to freeze–thaw cycles is one of the most important durability problems under subzero temperature conditions. It becomes more devastating when associated with marine environment. Supplementary cementitious material such as fly ash may be used as partial replacement of cement for making more dense, less absorptive and less permeable concrete which improve freeze–thaw durability. This paper presents a part of an experimental study on the freeze–thaw effect of concrete specimens exposed to artificial seawater simulating marine environment and plain water over 360 cycles. Three different grades of concrete M38, M33 and M28, each with four different fly ash replacement level, 20, 30, 40 and 60% were used for the experimental programme. The deteriorative effects were measured by studying weight and volume change, compressive strength, permeability characteristics and rapid chloride penetration resistance of the deteriorated test specimens. the optimum amount o...

Design of a Miniaturized Slotted T-Shaped Microstrip Patch Antenna to Detect and Localize Brain Tumor

Abstract: This paper presents a miniaturized, single-fed, slotted T-shaped, body worn antenna mainly designed for detecting and locating brain tumor which operates in the frequency band 902-928 MHz of Industrial, Scientific, and Medical (ISM) band. The key prominences of the proposed antenna are its minuscule dimension, vast bandwidth, good parameter results to identify affected and unaffected human brain tissue. The dimension of the proposed antenna is 29.99 mm × 29.99 mm × 0.59 mm, which has been placed over a complete human head phantom model consisting of six different layers; designed in CST Microwave Studio without altering their dielectric properties to carry out the simulation. Several performance measurements have been performed for both normal condition and tumor affected conditions changing the tumor positions with respect to the antenna. Analyzing these data, the location of the tumor can be estimated. A tumor having a radius of 5 mm with conductivity and permittivity of 7 S/m and 55 respectively, has been taken into consideration for the simulation process. The maximum SAR (specific absorption rate) of the proposed model measured 0.332 W/ Kg which satisfies the required safety guidelines.

Automatic Bengali Document Categorization Based on Word Embedding and Statistical Learning Approaches

Abstract: The automated categorization of text documents into predetermined categories has witnessed a growing in the last few years, due to the huge availability of documents in digital form and the ensuing need to organize them. Automatic document categorization is the process of assigning one or more categories or classes to a document, making it easier to manipulate and sort. This paper proposes a Bengali document categorization technique based on word2vec word embedding model and stochastic gradient descent (SGD) statistical learning algorithm with multi-class svm. The semantic features of a document are extracting by Word2Vec and SGD improve the classification complexity with multi-class SVM that classify the unlabeled data. The experimental result with 10000 training and 4651 testing documents shows the 93.33% accuracy.

R 3 HMS, An IoT Based Approach for Patient Health Monitoring

Abstract: Present tradition of Internet of Things (IoT) employed in healthcare services are not suitable in developing countries like Bangladesh due to the complexity, technical difficulty in maintenance and astronomical price tag. In order to fill this gap, this paper presents R3HMS(Remote Reliable and Real-time Health Monitoring System), an approach a to low cost, easy to maintain and easy to install IoT-based health monitoring system for 24 hours continuous screening of patients which can be employed in both large hospitals in urban area as well as in resource limited hospitals in rural areas. The system measures patients physiological signals like ECG, respiratory airflow and SpO2 and sends these extracted data to a remote server. ESP8266 Wi-Fi chip is chosen for wireless communication and an 8-bit MCU ATmega328p is used to accumulate and process data to reduce load on the ESP8266. With the help of Amazon Web Services(AWS)data is securely transferred between devices using MQTT messaging protocol. These data can be viewed and analyzed by specialists, doctors and other paramedical staffs in real-time to monitor patient’s state.

Design and Development of a Low-Cost IoT based Environmental Pollution Monitoring System

Abstract: Environment pollution is one of the foremost and crucial factors affecting lifestyles and health of human, living organisms, natural or built environment. This paper will provide a low cost environmental pollution monitoring system for monitoring the highly toxic gases like (CO 2 ,CO,&CH 4 ), sound pollution, temperature and dust measurements which is designed by observing users’ requirement to distinguish and keep away from disclosure to air and noise pollutants. The system consists of multiple gas sensors, dust device, sound sensor and temperature sensor that are integrated into a single platform. This paper will present a brand new system that includes digital hardware for obtaining atmospheric data and software to analyze the results. The designed system can provide calibration of the sensors along with improving the optimization to preserve vitality which results in the improvement of the precision of sensor information. In addition to this, the system is connected to the web via Wi-Fi or customer’s tablet or smartphone and can be interfaced to another device for the Internet of Things (IoT) based applications. This work makes an attempt to observe the quality of air and sound of a particular location and send this information to responsible people who will use this information to upgrade the standard of living of local people of that location, which is one of the first ideas about Smart City. The cost and effort of implementation is incredibly cheap and easy as the sensors and the microcontrollers are available in abundant. The online data storage system has extended the user friendly environment by modernizing all existing factors over a general server. The experimental results demonstrate the effectualness of this research work in terms of quick detection and real time response.

Design and Implementation of a Wireless Health Monitoring System for Remotely Located Patients

Abstract: Remote health monitoring is regarded as a hot topic in the arena of research. Though the number of aged people is increasing, it is irrefutable that the need of a disperse medical care system providing remote monitoring intending to reduce the escalating healthcare expenditure is very urgent. Moreover, World Health organization reported, in Bangladesh, about 17% of deaths are caused by cardiovascular diseases. Continual health monitoring and rapid detection can save up to 60% of lives. For these reasons, a wireless, wearable, low cost and automatic health monitoring system is a suitable solution. The implemented project described in this paper is a user- friendly, wearable and continuous monitoring system of blood pressure, heart rate, and body temperature, which can communicate with requested Android devices through Bluetooth. The system mainly consists of IR transmitter, receiver, LM-35, MPXV 5050GP, the data acquisition unit, microcontroller (i.e., Arduino) and Bluetooth. Bluetooth is used for its better transmission rate than ZigBee and cost-efficiency than GSM. In addition, all the measured readings and communications are password protected. A free and user- friendly Android application is attributed to demonstrate the readings. The proposed system has been field tested to ensure dependability and exactitude. The test results showed that this system could measure the patient‘s physiological data with high accuracy. The system can be easily segregated from the previous works for its maximum number of data collection, real-time monitoring, security issues and cost-effectiveness.

Highly birefringent large negative dispersion-flattened photonic crystal fibre for broadband residual dispersion compensation

Abstract: A triangular lattice photonic crystal fibre is presented in this paper for residual dispersion compensation. The fibre exhibits a flattened negative dispersion of −992.01 ± 6.93 ps/(nm-km) ...

Volumetric and Ultrasonic Velocity Studies of Urea and Thiourea in Aqueous Solution

Abstract: The observations on the anomalous behavior of urea and the comparison between urea and thiourea in aqueous solutions have been examined by volumetric and ultrasonic sound velocity techniques at different temperature (298.15, 303.15, 308.15, 313.15, 318.15 and 323.15 K), atmospheric pressure by using a high accuracy vibrating U-tube digital density and ultrasonic sound velocity analyzer. The apparent molar volume (φv) & apparent molar adiabatic compressibility (φk) have been calculated from experimental density and ultrasonic sound velocity data respectively and limiting apparent molar volume (φ0), limiting apparent molar adiabatic compressibility (φ k 0) have been evaluated from apparent molar volume vs. molality plot as intercept. Apparent molar expansibility (φE) was determined from apparent molar volume and hydration number (nH) from adiabatic compressibility. The results show very interesting information about strong solute-solvent & solute-solute interactions, and also elaborate the structure making or breaking behavior in the solution mixtures. keywords: Urea, Thiourea, Apparent molar volume, Expansibility, Sound velocity, Hydration number.

IoT Based Automatic Electricity Monitoring and Remote Load Control System Using PIC18F4550

Abstract: With the development in modern communication technology, every physical device is now connecting with the internet. IoT is getting emerging technology for connecting physical devices with the user. In this paper we combined existing energy meter with the IoT technology. By implementation of IoT in the case of meter reading for electricity can give customer relief in using electrical energy. In this work a digital energy meter is connected with cloud server via IoT device. It sends the amount of consumed energy of connected customer to webserver. There is a feature for disconnection in the case of unauthorized and unpaid consumption and also have option for renew the connection by paying bill online. We tried to build up a consumer and business friendly system.

Detection & Classification of Tumor Cells from Bone MR Imagery Using Connected Component Analysis & Neural Network

Abstract: Bone cancer is a class of diseases that are characterized by an unfettered growth of the cell and it is considered to be the main reasons of early death around the globe. Therefore, early detection and classification of the bone tumor are become needed to cure the patient. This study uses a connected component labeling algorithm for the detection of the bone tumor. In this work, the artificial neural network (ANN) is used for the classification of bone tumor. Total 220 bone MR images of previously verified patients are collected and the texture features of this images are used for the training and testing of the neural network. The obtained performance of the classification result exhibit that the neural network provides 92.50% success rate in bone tumor classification.

K-Nearest Neighbor Learning based Diabetes Mellitus Prediction and Analysis for eHealth Services

Abstract: Nowadays, eHealth service has become a booming area, which refers to computer-based health care and information delivery to improve health service locally, regionally and worldwide. An effective disease risk prediction model by analyzing electronic health data benefits not only to care a patient but also to provide services through the corresponding data-driven eHealth systems. In this paper, we particularly focus on predicting and analysing diabetes mellitus, an increasingly prevalent chronic disease that refers to a group of metabolic disorders characterized by a high blood sugar level over a prolonged period of time. K-Nearest Neighbor (KNN) is one of the most popular and simplest machine learning techniques to build such a disease risk prediction model utilizing relevant health data. In order to achieve our goal, we present an optimal KNearest Neighbor (Opt-KNN) learning based prediction model based on patient’s habitual attributes in various dimensions. This approach determines the optimal number of neighbors with low error rate for providing better prediction outcome in the resultant model. The effectiveness of this machine learning eHealth model is examined by conducting experiments on the real-world diabetes mellitus data collected from medical hospitals.