http://hdl.handle.net/20.500.12323/644 2026-09-12T15:30:31Z http://hdl.handle.net/20.500.12323/8347 Title: Twitter Social Network Data Mining in Healthcare Authors: Huseynova, Humay Abstract: Social media offers a wide range of opportunities to share information. In recent years transparency and open sharing via social media have significantly increased, hence, social media and healthcare became of interest among researchers, physicians, and doctors. Twitter is playing a significant role among other social media types because of its usability, so we can say that its use in healthcare is undeniable. It means that it is possible to see how much Twitter is capable to make sense in solving health problems. The main purpose of my research is to analyze in what capacity twitter is useful for healthcare, diseases, and treatments by using needed twitter data. I am going to use twitter mining techniques with simple Python codes to analyze tweets, to study the nature of sharing public health information on Twitter and to make a prediction for Azerbaijan environment by getting a reasonable conclusion. Description: School of Science and Engineering Major: 266147 - Software of Computer Systems and Networks Supervisor: Dr. Mahammad Sharifov Advisor: Behnam Kiani 2019-01-01T00:00:00Z http://hdl.handle.net/20.500.12323/6152 Title: The impact of security and privacy issues on data management in fog Computing Authors: Adepoju, Alamu Luke Abstract: With the increased growth of the application domains of IoT and the associated volumes of data generation, IoT systems are complicated and have small storage and recycling capacity. The cloud, a primary IoT storage medium with countless benefits, is not ideal for processing real time IoT data without delays. Capacity of data generated by IoTs keep increasing rampantly with associated security risks and privacy-preserving problems. Therefore, privacy maintenance, confidentiality and integrity of user’s data, improved latency and bandwidth restrictions are some of the major respective challenges of cloud computing. Fog computing is therefore a novel paradigm and an extension of the cloud. Which aims to improve cloud efficiency by enabling IoTs to locally process data before cloud transmission. However, some of the issues present in cloud such as the establishment of connection between edge devices often raise security and privacy concerns are also inherent in fog. The goal of this study, however, is to look at the state of data management security and privacy in a fog computing environment by reviewing existing security frameworks and data privacy procedures. This study lays bare the security vulnerabilities that exist inside the fog environment, creating hazards to user data privacy and security, and in lieu of that, this study incorporates features of data in addition to the acquired facts and statistics. Privacy-preservation is key to the continued use of services within the context of internet usage, as a result respondents indicated that they were experienced internet users who have been using the internet and its associated resources for various purposes, however respondents neither agreed nor disagreed with the possibility of the tracking or monitoring of their usage of the internet. The perception of respondents influenced the usage of the internet and various computing devices. 2022-05-01T00:00:00Z http://hdl.handle.net/20.500.12323/5484 Title: Generative Adversarial Network Image synthesis method for skin lesion Generation and classification Authors: Mutepfe, Freedom Abstract: Skin cancer is the most commonly diagnosed cancer in today's growing population. One of the common limitations in the treatment of cancer is in the early detection of this disease. Mostly, skin cancer is detected in its later stages, when it has already compromised most of the skin area. Early detection of skin cancer is of utmost importance in increasing the chances for successful treatment, thus reducing mortality and morbidity. Currently, most dermatologists use a special microscope to examine the pattern and the affected area. This method is time-consuming and is prone to human errors, so there is a need for detecting skin cancer automatically. In this study, we investigate the automated classification of skin cancer using the Deep Convolution Generative Adversarial Network(DCGAN).In this work, Deep Convolutional GAN is used to generate realistic synthetic dermoscopic images, in a way that could enhance the classification performance in a large dataset and to evaluate whether the classification accuracy is enhanced or not, by generating a substantial amount of new skin lesion images. The DCGAN is trained using images generated by the Generator and then tweaked using the actual images and allow the Discriminator to make a distinction between fake and real images. The DCGAN might need slightly more fine-tuning to ripe a better return. Hyperparameter optimization can be utilized for selecting the best-performed hyperparameter combinations and several network hyperparameters, namely number of iterations, batch size, and Learning rate can be tweaked, for example in this work we decreased the learning rate from the default 0.001 to 0.0002 and the momentum for Adam optimization algorithm from 0.9 to 0.5, in trying to reduce the instability issues related to GAN models. Moreover, at each iteration in the course of the training process, the weights of the discriminative and generative network are updated to balance the loss between them. This pretraining and fine-tuning process is substantial for the model performance. 2021-01-01T00:00:00Z http://hdl.handle.net/20.500.12323/5483 Title: Evaluation Of EOR Potential in Shale Oil Reservoir, Focusing on Wettability, and In-Situ Fluid Composition Authors: Ahmed, Imtiaz Abstract: Developing shale oil and gas resources is becoming essential due to the continuous depletion of conventional reservoirs. As the thrust towards shale oil resources increases, the petroleum industry, especially in countries striving to mitigate the challenges of such reservoirs. One of the essential techniques utilized to increase the production capability of such reservoirs is hydraulic fracturing. Besides, EOR gas injection assists the recovery of the process by increasing pressure and decreasing the oil's viscosity. This research evaluates EOR potential by focusing on wettability variations on recovery performance. In unconventional shale oil reservoir. The candidate reservoir is based on a simple layer cake model, simulated on a dual permeability approach with four fractured layers. The reservoir is first perforated, and then the reservoir is undergone through the EOR CO2 cyclic gas injection process. They huff and puff cycle has been done in the reservoir model for ten years. The research studies the effect of wettability on an unconventional reservoir. Three different wettability cases in 3 different permeability models, i.e., 0.00001 mD, 0.0001 mD, and 0.001 mD, are studied. The sensitivity result shows that the 0.001 mD model possesses the highest cumulative production accounts for 550,000 BBL, followed by 0.0001 mD and 0.00001 mD. Through sensitivity analysis and comparison, it has been concluded that wettability variations do affect the recovery performance in the reservoir model that is even below 0.01 mD permeability. Besides, the changes in the wettability in 3 different permeability distribution models significantly improve the production performance of the reservoir. 2021-01-01T00:00:00Z