Browsing by Author "Islam, Md. Milon"

Now showing 1 - 3 of 3

An Approach for Demand Forecasting in Steel Industries Using Ensemble Learning
(Daffodil International University, 2022-02-25) Raju, S. M. Taslim Uddin; Sarker, Amlan; Islam, Md. Milon; Al-Rakhami, Mabrook S.; Al-Amri, Atif M.; Mohiuddin, Tasniah; Albogamy, Fahad R.
This paper aims to introduce a robust framework for forecasting demand, including data preprocessing, data transformation and standardization, feature selection, cross-validation, and regression ensemble framework. Bagging (random forest regression (RFR)), boosting (gradient boosting regression (GBR) and extreme gradient boosting regression (XGBR)), and stacking (STACK) are employed as ensemble models. Different machine learning (ML) approaches, including support vector regression (SVR), extreme learning machine (ELM), and multilayer perceptron neural network (MLP), are adopted as reference models. In order to maximize the determination coefficient () value and reduce the root mean square error (RMSE), hyperparameters are set using the grid search method. Using a steel industry dataset, all tests are carried out under identical experimental conditions. In this context, STACK1 (ELM + GBR + XGBR-SVR) and STACK2 (ELM + GBR + XGBR-LASSO) models provided better performance than other models. The highest accuracies of R2 of 0.97 and 0.97 are obtained using STACK1 and STACK2, respectively. Moreover, the rank according to performances is STACK1, STACK2, XGBR, GBR, RFR, MLP, ELM, and SVR. As it improves the performance of models and reduces the risk of decision-making, the ensemble method can be used to forecast the demand in a steel industry one month ahead.
Analysis and Modeling of Automated Walking Guide to Enhance the Mobility of Visually Impaired People
(Khulna University of Engineering & Technology (KUET), Khulna, Bangladesh, 2019-03) Islam, Md. Milon; Sadi, Prof. Dr. Muhammad Sheikh
The development of walking guides has become a prominent research due to the rapid growth of visually impaired people in recent decades. Although numerous systems have been developed to aid the visually impaired people, a considerable portion of these are limited in their scopes. This thesis has implemented a spectacle prototype to assist these individuals with safe and efficient walking in the surrounding’s environment. The spectacle prototype is modeled in SolidWorks (3D model) considering the dimension of each electronic components. In the modeling, the front ultrasonic sensor is positioned in the spectacle to detect the front obstacles only, the left and right ultrasonic sensors are set to 45 degree from the spectacle center point in order to detect obstacles within the shoulder and arm of user; another ultrasonic sensor is positioned towards the ground facing for the detection of pothole. The Rpi camera is positioned at the center point of the spectacle. In addition, the right and left temple of the spectacle is designed to position the raspberry pi and battery respectively. The usage of spectacle based walking guide would help the visually impaired people to scan the surroundings. Three pieces of distance measurement sensors (ultrasonic sensor) is used in the walking guide in order to detect the obstacle in each direction including front, left and right. In addition, the system detects the potholes on the road surface using sensor and convolutional neural network (CNN). Overall, the spectacle prototype consists of four ultrasonic sensors; raspberry pi, Rpi camera and battery. CNN technique, runs on raspberry pi, is used to detect the pothole on the road surface. The pothole images are trained initially using convolutional neural network in a host computer and the potholes are detected by capturing a single image each time. The experimental study demonstrates that 98.73% accuracy is achieved by the front sensor with an error rate of 1.26% when the obstacle is at 50 cm distance. In addition, the results reveal that the system obtains the highest accuracy, precision and recall 92.67%, 92.33% and 93% respectively for potholes detection. The electronic spectacle gives a direct audio signal to the user via headphone for avoiding hindrances effectively.
Performance Evaluation of Random Forests and Artificial Neural Networks for the Classification of Liver Disorder
(IEEE, 2018-09-20) Haque, Md. Rezwanul; Islam, Md. Milon; Iqbal, Hasib; Reza, Md. Sumon; Hasan, Md. Kamrul
Liver is the major organ inside the human body which is very supportive for digesting food, eliminating poisons, and stocking energy. The rate of Liver disorder patients is rapidly rising all over the world. But it is very hard to identify the disorder from its ambiguous symptoms which increases the mortality rate due to this disease. The paper represents an expert scheme for the classification of liver disorder using Random Forests (RFs) and Artificial Neural Networks (ANNs). The methods train the input features using 10-fold cross validation fashion. The dataset named as BUPA liver dataset is retrieved from UCI machine learning repository for our research study. The performance of the proposed scheme is assessed in view of accuracy, positive predictive value, negative predictive value, sensitivity, specificity and F1 score. The scheme delivers a better result for training but comparatively low for testing. The scheme obtained the accuracy of 80% and 85.29% by RFs and ANNs respectively along with the F1 score of 75.86% and 82.76% in testing phase.