Turkish Journal of Computer and Mathematics Education (TURCOMAT)

A Review on Breast Cancer Prediction Using Machine Learning and Deep Learning Techniques

Mounika potta, B. Narayanan, Kavitha Rani Balmuri — Wed, 04 Sep 2024 00:00:00 +0000

Breast cancer is one of the most prevalent and chronic disease that affect women. To overcome this disease, effective medical treatment is required. Early detection of the disease plays an important role for suitable medication and survival of patient. To identify the breast cancer in the patients, standard imaging technique mammography is used. Due to the subtle and varied nature of cancer tissues interpreting mammogram images can be a challenge to doctors. Machine learning (ML) and Deep Learning (DL) techniques offer promising solutions that provide efficient breast cancer detection from mammograms. In this review paper a comprehensive review of ML and DL algorithms and their applications in mammogram image analysis are presented. Various supervised and unsupervised learning techniques, such as convolutional neural networks (CNNs), support vector machines (SVMs), random forests, and other popular ML and DL models are discussed in paper. The integration of these DL methods that are efficiently used in image preprocessing techniques, feature extraction, and classification strategies. The overall survey focusses on various performance metrics, datasets, and benchmarks used in existing studies. Further the strengths and limitations of different approaches used by various researchers are identified. By understating current research trends this paper aims to contribute to the ongoing development of more accurate and reliable breast cancer detection systems using advanced ML techniques.

ADVANCED AUTONOMOUS SURVEILLANCE ROBOT FOR ENHANCED MONITORING AND INDIVIDUAL IDENTIFICATION

Mon, 09 Sep 2024 00:00:00 +0000

The primary objective is to detect and identify suspicious activities and potential threats in a precise manner while prioritizing human safety leveraging surveillance technology and machine learning. The implementation of this system involves coding in Python using the OpenCV library. It utilizes Wi-Fi connectivity as a means of communication. The robot is equipped with a Raspberry Pi along with a USB web camera, which captures video footage and employs object detection algorithms to identify unknown individuals. When a person or an object is detected, the system sends an email to the dedicated email addresses including an image of the unrecognized individual. The proposed system is designed as a unified unit responsible for monitoring the environment for hazardous conditions and delivering real-time video feedback. The proposed system is simulated and tested in real-time to observe its functionality, and it is observed that the system works properly as per given input conditions.

A Secure Crypto-Biometric System Utilizing GMM Encoder and BCH

Ms. Sanjeevini , Ashraya B. , Aparnika G. , Akshitha M. — Wed, 04 Sep 2024 00:00:00 +0000

Now that cloud computing has reached maturity, a diverse array of providers and services are available in the cloud. On the other hand, security issues continue to receive a lot of focus. Despite the many benefits of cloud computing, users are hesitant to embrace the technology due to concerns about their security and privacy. While biometric technologies are rapidly becoming an integral part of many secure identification and personal verification solutions, they do pose certain challenges when stored in the cloud owing to privacy regulations and the requirement to have faith in cloud providers when handling biometric data. In this work, we offer a crypto biometric system that can be used with cloud computing to solve these issues. This system ensures that no private biometric data is revealed.

Botnet Attack Identification and Mitigation condition Software-Defined Networks Utilizing CNN Algorithm

Wed, 04 Sep 2024 00:00:00 +0000

One new design that makes managing and communicating across large-scale networks easier and more flexible is software-defined networking, or SDN. It allows for the smooth and dynamic execution of complicated network choices via programmable and centralized interfaces. But SDN opens doors for people and companies to tailor network apps to their needs, allowing them to enhance services. On the other hand, it began to encounter a host of new privacy and security issues and brought the dangers of one point of failure all at once. In most cases, hackers use OpenFlow switches to conduct botnets or distributed Denial of Service (DDoS) assaults against the controller. Popular security apps that use deep learning (DL) to quickly identify and counteract attacks are on the rise. Here, we examine botnet-based DDoS attack detection using DL approaches in an SDN-supported context and demonstrate their performance. For the assessment, we utilize a dataset that we just created ourselves. In order to choose the most useful subset of characteristics, we used weighting of features and tuning techniques. Using both a synthetic dataset and actual testbed conditions, we validate the measurements or simulation results. The primary objective of this research is to identify botnet-based DDoS assaults using easily-obtained characteristics and data using a lightweight DL approach with baseline hyper-parameters. We found that the DL technique's performance is affected by the optimal subset of features, and that the accuracy of predictions of the same approach may be varied with a different collection of features. Lastly, our empirical findings show that the CNN approach works better than both the dataset and the actual testbed environments. With CNN, the detection rate for typical flows is 99% and for malicious flows it drops to 97%.

An approach for two-dimensional convolutional neural networks for hourly passenger boarding demand prediction based on uneven smart-card data

Wed, 04 Sep 2024 00:00:00 +0000

An invaluable resource for understanding passenger boarding patterns and forecasting future travel demand is the tap-on smart-card data. Positive instances, on the other hand—boarding at a given bus stop at a certain time—are less common than negative instances when looking at the smart-card data (or instances) by boarding stops and by time of day. Machine learning algorithms that are used to estimate hourly boarding numbers at a certain location have been shown to be much less accurate when the data is imbalanced. Before using the smart-card data to forecast bus boarding demand, this research tackles the problem of data imbalance in the data. To create fake traveling instances to add into a synthetic training dataset containing more evenly distributed traveling and non-traveling examples, we suggest using deep generative adversarial networks (Deep-GAN). Next, a deep neural network, or DNN, is trained on the synthetic dataset to predict which instances from a given stop in a certain time frame will travel and which ones won't. According to the findings, resolving the data imbalance problem may greatly enhance the predictive model's functionality and make it more accurate in predicting ridership profiles. The suggested strategy may create a synthetic training set with a better similarity so diversity and, therefore, a stronger prediction capability, according to a comparison of the Deep-GAN's performance with other conventional resampling techniques. The study emphasizes the importance of the issue and offers helpful recommendations for enhancing the quality of the data and model performance for individual travel behavior analysis and travel behavior prediction.

Spam Review Detection Using Weighted Swarm Support Vector Machines and Pre-Trained Word Embedding for Multiple Languages

Mr. D. Shine Rajesh , Pravalika K. , Nishika B., Joshitha Sree D. — Wed, 04 Sep 2024 00:00:00 +0000

Before making a purchase, many find internet reviews to be a vital source of information. On top of that, companies may learn a lot about their products and services via these reviews. Having faith in these reviews was especially important during the COVID-19 pandemic, when many stayed inside and read reviews at a dizzying pace. The pandemic altered the atmosphere and people's preferences in addition to increasing the number of evaluations. Spam reviewers keep an eye on these changes and try to improve their sneaky techniques. In order to deceive customers or harm competitors, reviews that are deemed spam may include inaccurate, misleading, or dishonest information. Consequently, this work introduces a WSVM plus an HHO to identify spam reviews. The HHO is similar to an algorithm in that it optimises hyperparameters and uses feature weights. Using English, Spanish, and Arabic language corpora as datasets, the multilingual difficulty in spam reviews has been tackled. Ngram-3, TFIDF, whereas One-hot encoding are three methods for representing words, while pre-trained word incorporation (BERT) is another one that has been used. Each of the four such studies has shed light on and provided a solution to a different facet. From start to finish, the proposed technique beat rival cutting-edge algorithms in every test. For the Multi dataset, the WSVM-HHO achieved a success rate of 84.270 percent; for the English information set, 89.565 percent; for the Spanish information set, 71.913 percent; and for the Arabic dataset, 88.565 percent. Furthermore, we have extensively researched the review environment before to and during the COVID-19 event. To further enhance detection performance, it has been designed to merge its existing textual attributes with statistical information to build a new dataset.

Eye Deep-Net: a deep neural network-based multi-class retinal disease diagnostic

Mr. Anil P Jawalkar , Navya D., Himasriya E., Harika sree Y. — Wed, 04 Sep 2024 00:00:00 +0000

Ophthalmologists rely heavily on retinal pictures to diagnose a wide range of eye conditions. Numerous retinal disorders may lead to microvascular alterations in the retina, and a number of studies have been conducted on the early identification of medical pictures to enable prompt and appropriate treatment. In order to identify various eye illnesses using color fundus pictures, this study develops a non-invasive, automated deep learning system. A multiclass ocular illness A productive diagnostic approach was created using the Remind dataset. A variety of augmentation strategies were used to make the structure robust in real-time after multi-class fundus pictures were collected from a multi-label dataset. Low computational demand images were processed in accordance with the network. The fundamental convolutional neural network (CNN) extracts appropriate characteristics from the input color fundus image dataset, and then processed characteristics were employed to make predictive diagnoses. This multi-layer neural network, called Eye Deep-Net, has been developed for training and evaluating images for the recognition of various eye problems. The performance of the suggested model is determined to be much better than numerous baseline state-of-the-art models. The strength from the Eye Deep-Net is assessed using different statistical metrics. The suggested methodology's effectiveness in classifying and identifying diseases using digital fundus pictures is shown by a thorough comparison with the most modern techniques.

Price Negotiating Chatbot on E-commerce website

Mr. Bhanu prasad Gorantla , Pravallika G. , Ashritha K. , Sathwika K. — Wed, 04 Sep 2024 00:00:00 +0000

The rise of internet purchasing in the last few years is quite remarkable. Despite this growth, not all aspects of internet buying have been perfected. For example, unlike in physical stores, you can't haggle with vendors about prices. A chatbot for product negotiations is now live. Customers are able to acquire a good deal on product(s) with the help of the chatbot. The approach might end up hurting either the goods seller or the customer's budget, as it affects a lot of different parts of online buying. We have devised an algorithm that, in conjunction with the forecast of previously accessible data, can offer a price in order to circumvent such scenarios. Using unrelated data elements or qualities or techniques that aren't a good fit for a certain dataset might reduce the accuracy of price prediction. In light of the fact that erroneous product price predictions may lead to significant financial losses for online retailers, these companies avoid relying only on price prediction algorithms. When data becomes too large or when a characteristic that was relied on the model's prediction becomes unavailable, certain models can fail. Then, in order to keep the model's accuracy and dependability intact, such modifications must be handled. We have made an effort to address some of these concerns in our chatbot system.

Autonomous Road Damage Detection using Unmanned Aerial Vehicle Images and YOLO V8 Methods

Wed, 04 Sep 2024 00:00:00 +0000

Using photos from Unmanned Aerial Vehicles (UAVs) and deep learning algorithms, this research provides a revolutionary automated road damage identification method. In order to provide a secure and long-lasting transportation system, road infrastructure maintenance is essential. On the other hand, gathering road damage data by hand may be dangerous and labor-intensive. Therefore, we suggest using artificial intelligence (AI) and unmanned aerial vehicles (UAVs) to greatly increase the effectiveness and precision of road damage identification. For object recognition and localisation in UAV photos, our suggested method makes use of three algorithms: YOLOv4, YOLOv5, and YOLOv7. We used a mix of a Spanish roadway dataset and the Chinese RDD2022 dataset for training and testing these methods. Our method obtains 59.9% average precision (mAP@.5) for the YOLOv5 versions, 65.70% mAP@.5 when using the YOLOv5 version using the Transformers Prediction the Head, or 73.20% mAP@.5 for that YOLOv7 version, testing results show the effectiveness of our methodology. These findings open the door for further study in this area and show the possibilities of employing deep learning and UAVs for automatic road damage identification.

Using CNN, GRU, and B/idirectional Multiscale Convolutional Neural Networks for Human Behavior Recognition

Mrs. Naga Lakshmi , Rashmi M., Sathvika M., Trishika V. — Wed, 04 Sep 2024 00:00:00 +0000

The main challenge in recognizing human behavior is constructing a network for the extraction and categorization of spatiotemporal features. In order to address the issue that the current channel attention mechanism simply aggregates each channel's global average information while ignoring its specific spatial information, this work suggests two enhanced channel attention modules: the depth separable convolutions section and the time-space (ST) interaction section of matrices operation. These modules are also combined with research on the recognition of human behavior. Proposing a multiple habitats convolutional neural network technique for human behavior detection, it is combined with the excellent performance using convolutional neural network (CNN) for video and image processing. First, the behavior video is divided into segments. Next, low rank learning is applied to each segment to extract the associated low rank actions information. Finally, these minimal position behavior information are linked together in the time axis to get the low are behavior data for the entire video. This allows for the efficient extraction of behavior information from the video without the need for laborious extraction processes or assumptions. Neural networks can simulate human behavior in a variety of network topologies by transferring and reusing this capacity. To lessen the distinction between features derived from various network topologies, two efficient feature difference measurement methods are presented, taking into account the various properties of data features at various network levels. The suggested strategy has a decent categorization impact, according to experiments on a number of available datasets. The experimental findings demonstrate that the method's accuracy in identifying human behavior is excellent. It has been shown that the suggested model increases recognition accuracy while simultaneously enhancing the compactness for the model structure and successfully lowering the computational cost of the output weights.

An Innovative Hybrid Approach to Forecasting Soluble Oxygen for Optimal Water Purification in Highly Concentrated Aquaculture

Mr. Shaik Mohammed Imran , Yashaswini D., Sahasra Ch., Gayathri B. — Wed, 04 Sep 2024 00:00:00 +0000

An important measure of the water's quality in an aquaculture setting is the concentration of dissolved oxygen. Disintegrating oxygen content prediction using conventional techniques is slow and inaccurate due to the complexity, nonlinearity, and dynamics of the process. This research develops a hybrid model that addresses these problems by combining the radiation gradient enhancement machine (LightGBM) with this simple rechargeable unit (Biru). The first step was to find the important parameters by using linear interpolation and smoothing. After removing superfluous variables, the LightGBM algorithm predicts dissolved oxygen in highly intensive aquaculture and establishes its relevance. Lastly, the attention approach was used to map the learning parameter matrices and weighting matrices, allowing various weights to be applied to the Biru's hidden states. The results show that the given prediction model can capture the upward trend of oxygen dissolution fluctuations over a 10-day period with a rate of accuracy reaching 96.28% in only 122 seconds. It takes the least amount of time to compare the model impacts of Biru-AAttention, LightGBM-GGRU, LightGBM-LSTM, as well as LightGBM-BBiru. The improved accuracy of its predictions makes it a valuable tool for controlling the water quality in intensive aquaculture.

Comprehensive Data Corruption Identification Using Machine Learning Algorithms (PAACDA)

Dr. M. Vanitha , Maneesha K. , Uma Renu Sri K. , Nancy K. — Wed, 04 Sep 2024 00:00:00 +0000

Data and analysis have evolved from being scattered numbers and qualities in spreadsheets to being seen as a means to revolutionize any substantial industry, thanks to the rise of technology. There are many unethical and unlawful ways that data may get corrupted; thus, it's important to find a way to effectively detect and highlight all the corrupted data in the dataset. It is not an easy task to detect damaged data or to restore information from a corrupted dataset. This is crucial and could cause issues with data processing using machines or deep learning methods later on if not handled early enough. Rather than focusing on outlier identification, this study introduces its PAACDA: Presence-driven Adamic Adar Corruption Identification Algorithm and then consolidates the findings. Even though they rely on parameter tuning to achieve high accuracy, and remember, current state-of-the-art models like Isolation Forest and DBSCAN (which stands for "Density-Based the Spatial Process of Clustering of the Applications with Noise") have a lot of uncertainty when they factor in corrupted data. This study investigates the specific performance problems with several unsupervised learning methods on corrupted linear and clustered datasets. In addition, we provide a new PAACDA technique that achieves a higher precision of 96.35% for cluster data and 99.04% for linear data compared to previous unsupervised training benchmarks on 15 prominent baselines, including K-means clustering, isolation forest, and LOF (local outlier factor). From the aforementioned angles, this essay delves deeply into the relevant literature as well. In this study, we identify all the problems with current methods and suggest ways forward for research in this area.

Memory Loss and Alzheimer's Disease Progression Convolutional Neural Networks with Dropout Layers for Optimal Filtered Features in MRI Images of the Hippocampus for Slice Selection Based on Landmarks

Wed, 04 Sep 2024 00:00:00 +0000

The public health threat of Alzheimer's disease (AD) is now widely accepted. When using machine learning techniques and MRI scanning to detect Alzheimer's disease, the hippocampi are readily accessible and one among the most afflicted brain regions. AD classification by machine learning algorithms using complete MRI slices was unsatisfactory. This article describes how to choose MRI slices using hippocampus landmarks. This research aims to find the best accurate AD categorization MRI pictures. Next, utilizing Resnet50 or LeNet using various classifiers with the open-source and free ADNI dataset, the three views and categories were valued. The models used 4,500 Neuroimaging slices from three perspectives and categories for training. We found that AD classification was better with MRI scan segments than whole slices. The coronal view showed our method's machine learning accuracy enhancement most clearly. This strategy greatly enhanced machine learning accuracy. The findings from a rotational perspective matched what clinicians use to identify AD. Additionally, LeNet models may classify AD effectively.

Machine Learning for Cloud-Based Privilege Escalation Attack Detection and Mitigation with CATBOOST

Dr.AR.Sivakumaran, Sreeja G., Poojitha Ch., Ramya I. — Wed, 04 Sep 2024 00:00:00 +0000

The exponential growth in attack frequency and complexity in the past few years has made cybersecurity a major concern with the advent of smart devices. Cloud computing has changed the way businesses operate, but users may find it more challenging to use dispersed services, such as security systems, due to their centralization. Organizations and cloud service suppliers exchange massive amounts of data, which poses a significant risk of accidental or intentional disclosure of sensitive information. Because of their increased access and potential to do substantial harm, an antagonistic insider poses a serious threat to the company. Only approved individuals within the organization have access to sensitive data and assets. This research details a machine learning-based strategy for classifying insider threats and finding out-of-the-ordinary events that can indicate privilege escalation security issues. The system uses a systematic approach to detect these irregularities. Machine learning and prediction accuracy are both enhanced by ensemble learning, which considers several models simultaneously. Using anomaly and weakness detection, some studies have attempted to identify security issues or hazards associated with privilege delegation in network systems. However, the assaults cannot be definitely identified from this research. Ensembles for machine learning (ML) are suggested and assessed in this research. The objective of this endeavor is to classify insider assaults using machine learning approaches. The dataset it uses has been modified from many files beneath the CERT dataset. The dataset is subjected to four machine learning techniques: Light GBM, XG Boost, Ada Boost, and three Random Forest (RF) methods. In terms of overall performance, light was superior. In contrast, RF and AdaBoost are two algorithms that may be better at preventing assaults from inside, such as attacks using behavioral biometrics. Consequently, it is possible that various internal threats may be better classified by combining various machine learning algorithms. With a 97% dependability rate, the Light GBM method outperforms the other suggested techniques; RF, AdaBoost, and XG Boost all have 88% accuracy rates.

An Automated News Text Classification Information System

Mr.Munimanda Premchander, Are Deepthi, Bathula Swathi, Srija B. — Wed, 04 Sep 2024 00:00:00 +0000

An information system for the categorization of news texts using machine learning algorithms is being planned and developed in this project. An online platform and an automated categorization system make up the data system in question. We have preprocessed the text data. In order to train classifiers using the grid search method, many experiments were carried out. We have tested four different categorization algorithms: naïve Bayesian, logistic regression, random forest, and artificial neural network. Several measures, including F-score, recall, and precision, have been used to assess the trained classifiers' classification quality. An additional goal in developing the website was to provide easy access to the information system.

Darknet Traffic Analysis: Examining How the ADABOOST Algorithm Affects the Classification of Onion Service Traffic Given Modified Tor Traffic

Mr. Francis Vijay Kumar Anna Reddy , Yamini K., Kruthi K. , Akshitha Sree K. — Wed, 04 Sep 2024 00:00:00 +0000

In order to shape and monitor traffic, it is necessary to classify network traffic. The significance of privacy-preserving technology has increased in the last twenty years due to the growth of privacy concerns. One common method of remaining anonymous while surfing the web is to join the Tor network. This will allow you to remain anonymous while also supporting anonymous services called Onion Services. The problem is that government and law enforcement organizations often take advantage of this anonymity, particularly with Onion Services, and end up de-anonym zing its users. This paper's emphasis is on three primary contributions in an effort to discover the capability to categorize Onion Service traffic. Separating Onion Service communication from regular Tor traffic is our first objective. With over 99% accuracy, our methods can detect Onion Service traffic. On the other hand, Tor traffic may have its information leaking concealed by making a

Few adjustments. We assess the efficacy of our methods in light of these changes to Tor traffic in our second contribution. According to our experiments, under these circumstances, the Onion Services traffic becomes less distinct, with an accuracy decrease of over 15% seen in some instances. We conclude by determining and assessing the effect of the most important feature combinations on our classification task.

CNN2D Algorithm for Detection of Ransomware Attacks Using Processor and Disk Usage Data

Ms. Putta Srivani , Uma Sri J. , Reshmitha K., Soukya P. — Wed, 04 Sep 2024 00:00:00 +0000

Commonly, ransomware encrypts data, turns off antivirus protection, and destroys the target computer and everything on it. The techniques used today to detect this kind of WannaCry include monitoring the files, system requests, and processes on the system that is being targeted and analysing the data collected. Monitoring numerous processes has a substantial overhead; more current ransomware may interfere with the monitoring and alter the collected data. A dependable and practical technique for locating ransomware operating within a virtual machine, also called a VM, is provided in this study. We construct a framework for detection by applying an automated machine learning (ML) evaluation to the whole virtual machine (VM) using data collected from the physical host computer pertaining to specific processors and disc I/O events. This approach eliminates the need to continuously watch every action on the system that is being targeted and lessens the likelihood that ransomware would contaminate data. It also endures shifts in the amount of labour that users must do. It provides fast and very likely detection of known ransomware (used to train this machine learning model) and also of unknown ransomware (not utilised for teaching the model). Out of the seven artificial neural network classifiers that we looked at, the randomly generated forest (RF) classification gave the best results. Across six different customer loads plus 22 instances of ransomware, the RF model detected malware with a 0.98 confidence in 400 milliseconds.