Volume 14, Issue 1, February 2023

A Multifactor Analysis Model for Stock Market Prediction
Pages: 1-6 (6) | [Full Text] PDF (258K)
Akash Deep
Texas Tech University, Lubbock, TX 79409, USA

Abstract -
Stock Market predictions have historically been a problem tackled by different singular approaches even though markets are influenced by many different factors. This paper presents a novel multi-factor analysis model for stock price prediction that combines Technical analysis, Fundamental analysis, Machine learning, and, Sentiment Analysis (TFMS Analysis). The proposed model leverages Random Forest Regressor (RFR) to predict a stock price and long short-term memory(LSTM) approach to predict a multiplier. Sentiment analysis is then used to capture the impact of various factors on stock prices, including market trends, economic indicators, and public opinion. The results of the model are compared to traditional prediction models using historical stock data, and it is shown that the proposed model provides improved accuracy in predicting future stock prices. The proposed model represents a significant step forward in stock price prediction, providing a more comprehensive and effective approach to predicting stock prices based on multiple factors.
 Index Terms - Multifactor Analysis, Random Forest Regressor, LSTM, Sentiment Analysis, and Machine Learning Stock Price Prediction
Automatic Colorization and Detection of Tumor in Brain MRI Images Through Generative Adversarial Neural Networks
Pages: 7-18 (12) | [Full Text] PDF (865K)
Muhammad Usman, M.J. Arshad
Department of Computer Science, University of Engineering and Technology Lahore, Pakistan

Abstract -
Clinical image processing is an important part of today's healthcare systems and a commonly utilized approach. The 10th most prevalent type of tumor, the intracranial tumor, affects both children and adults. In medical imaging relating to intracranial tumors, there has been tremendous growth. Even when intracranial tumor pictures are appropriately captured, accurate tumor segmentation in the brain is difficult. The tumor can be cured if its stage and shape are detected early enough, and researchers have been developing advanced tests and approaches for this purpose. The MRI image collection is used in our T1-CE project. To discover irregularities in a medical picture that aid in diagnosis, an automated identification, segmentation, and colorization of tumor area is performed. Only the tumor location is colored in a grayscale picture to aid medical practitioners in visualizing tumor shape, size, and orientation. MRI images with color tumor areas were created using Cyclic GAN as well as with Pix2Pix Conditional Generative Adversarial Neural Networks (Pix2Pix- cGANs). In comparative analysis Pix2Pix network outperformed the Cycle Generative Adversarial network. The model's performance was evaluated using the Structure Similarity Index (SSIM) and the Peak Signal to Noise Ratio (PSNR). On generated images, we attained an average SSIM score of 0.92 percent and an average PSNR value of 28 percent, outperforming previous methodologies qualitatively.
 Index Terms - Tumor, Brain MRI, Segmentation, Colorization, CycleGAN and Pix2Pix GAN
Triple Block Data Security Based on Distributed Crypto-Steganography in the Cloud Environment
Pages: 19-25 (7) | [Full Text] PDF (578K)
Derrick M├ęthode BAGAZA, Blaise Omer YENKE, Jean M'BOLIGUIPA
Faculty of Science, University of Bangui, Central African Republic
Department of Computer Science, University Institute of Technology, University of Ngaoundere, Cameroon

Abstract -
Recent advances in cloud architecture led to the need to ensure more the privacy of data stored online. Data security is one of the most sensitive issues in Cloud environments. Indeed, users would like to have guarantees on the availability and integrity of their data. Multiple strategies have been deployed to achieve data security: making information completely unintelligible to unauthorised entities using cryptography, hiding the existence of information using steganography so that it cannot be detected, controlling access to information using blockchain technology and attribute-based encryption for secure interaction. Many authors have tried to reconcile steganography and cryptography by using ciphers or combinations of ciphers such as AES, RSA, Blowfish and the LSB technique. LSB has flaws due to the quality of additive noise which affects the statistical properties of the image. Furthermore, some authors have used the DCT technique, but DCT has limitations with respect to steganalysis when used alone. In this work, LSB and DCT techniques are combined for data integration and extraction to further enhance security. We call this combination LSB-DCT. Numerous experiments have been conducted to test the proposed approach and the results show a better data security performance index than the literature works.
 Index Terms - AES, Cloud Computing, Cryptography, LSB-DCT, RSA and Steganography
Secure and Versatile Decentralized Ledger System Based on Blockchain for P2P Communication
Pages: 26-32 (7) | [Full Text] PDF (444K)
Komal Shahzadi
Department of Computer Science and Engineering, University of Engineering and Technology, LHR

Abstract -
Data sharing between peers is also required, just like a secure data sharing infrastructure. Since, present systems have a centralized architecture that requires authentication from only a single authority, there have been a lot of privacy, security, and interoperability issues and security concerns have grown. Blockchain is considered to build the transparent, decentralized, and trustworthy systems. In this research, a framework has been proposed based on Blockchain technology offers a novel tool for solving information security, product traceability and efficiency, privacy, and cost reduction in peer-to-peer information exchange high security and cost-effectiveness for peer-to-peer data exchanging without the intervention of third-party. This framework has eliminated the need for third-party authentication systems, that ultimately save time and money. As a proof of concept, a use case of healthcare has been taken. There is an increasing necessity of immediate access to relevant healthcare information as in most of the cases there is need of data and information exchanging between different hospitals for better treatment of patients. Additionally, the proposed framework ensures the security of patients' health information using data encryption, enabling the establishment of a conditional access control system where only authorized parties are permitted to view patients' medical records. The cost and time analysis of proposed framework is also provided in this research.
 Index Terms - Smart Contract, Peer-To-Peer Communication, Blockchain, Ethereum and Electronic Health Record System
Pages: - (8) | [Full Text] PDF (395K)

Abstract -
 Index Terms -