Due to the advancement of Technologies, many companies and organizations today have huge databases, that grow to a limit of millions of records per day. This data generation and storage has become faster than ever. This resulted in databases of unbounded growth. The data even changes over time. As a consequence, if the data cannot fit into memory, it may be necessary to sample a smaller training set. The data streams have recently emerged to address the problems of continuous data. The mining of such data streams is a unique opportunity and even challenging task. Mining with data streams is the process of extracting knowledge structures from continuous, rapid data records. Due to increased streaming information, data stream mining has attracted the research community. Analyzing data streams helps in applications like scientific applications, business and astronomy etc. In this paper, we present the overview of the growing field of Data Streams. We cover the theoretical basis needed for analyzing the streams of data. We discuss the various techniques used for mining data streams. The focus of this paper is to study the problems involved in mining data streams. Finally, we conclude with a brief discussion of the big open problems in the area.
chronic kidney disease (CKD) is a perennial condition where the kidneys deteriorate and stop functioning gradually. This disease has become one of the major public health concerns worldwide. It is insidious, often recognizable only by laboratory abnormalities until its latest stages. The main motive of this work is to ascertain the existence of chronic kidney disease by imposing various classification algorithms on the patient medical record. This research work is primarily concentrated on finding the best suitable classification algorithm which can be used for the diagnosis of CKD based on the classification report and performance factors. Empirical work is performed on different algorithms like Support Vector Machine, Random Forest, XGBoost, Logistic Regression, Neural networks, Naive Bayes Classifier. The experimental results show that Random Forest and XGBoost give better results when compared to other classification algorithms and generates 99.29% accuracy.
19 January 2021 in International Journal of Online and Biomedical Engineering (iJOE)
Computed Tomography (CT) images are cross-sectional images of any specific area of a human body which allows doctors to see inside of a patient. CT scan is almost always the first imaging modality used to assess patients with suspected hemorrhage. A CT scan provides image reports in the form of grey shades. It is sometimes difficult to distinguish between two areas because the shades of grey in a CT image are occasionally similar. CT scan (Particularly “Non-Contrast Head CT Scan”) is the current guideline for primary imaging of patients with any head injuries or brain stroke like symptoms. To obtain any findings from the CT image, Radiologists or other doctors need to examine the images. Deep-learning is an important tool used in radiology and medical imaging which provides a better understanding of the image with more efficiency and quicker exam time. The main idea of this project is developing a model using classification algorithms which can be used to classify or detect hemorrhage in a CT image. The dataset consists of both normal CTs and CTs with hemorrhage. Deep learning is used to develop a model that can detect whether a CT image shows a hemorrhage or not.
7 July 2023 in International Journal of Online and Biomedical Engineering (iJOE)
Colors are the smiles of nature. Digital image processing has widely employed the use of the function the color due to its effectiveness as a tool in classifying and identifying objects, which can be distinguished based on various relevant shades of color. This work aims to create a software tool that assists color-blind individuals in identifying the colors and edges in an image that may appear similar to them. If you are not suffering from a color vision deficiency it is very hard to imagine how it looks like to be color blind. Individuals with color blindness are limited and sometimes even disqualified from specific professions due to their inability to differentiate between colors. The objective of this research is to create a technique or approach for accurately identifying different shades of colors and predicting their names specifically for individuals with color blindness. This project utilizes image processing techniques to create a system that can recognize and distinguish colors in an image. The algorithm used for color detection in this project recognizes pixels in an image that correspond to a defined color or color range and provides the name of the color along with its corresponding B, G and R values. This work will be very helpful for people who suffer from color blindness especially when they want to know the color of the things.
Associative Classification is a recent and rewarding approach which combines associative rule mining and classification. This technique has attracted many researchers as it derives accurate classifier with effective rules. Associative classifiers are useful for application where maximum predictive accuracy is desired. Healthcare industry collects large amounts of data which are not mined to discover hidden information for taking decision. Increasing access to huge datasets and corresponding demands to analyze these data has led to the development of new online algorithms for performing machine learning on data streams. There is a need to develop a decision support system for predicting the huge datasets generated by healthcare industry. In this paper we proposed an efficient technique for heart disease prediction. This research uses associative classification which builds a classifier with prediction rules of high interestingness values. Experimental results show that this work helps doctors in their diagnosis decisions.