Статьи журнала - International Journal of Mathematical Sciences and Computing
Все статьи: 240
Статья научная
Stroke prediction is paramount in healthcare to enable timely intervention and reduce the burden of this devastating condition. This research paper examines the prediction of strokes using machine learning methods, aiming to enhance accuracy and efficiency in risk assessment. Numerous Machine Learning (ML) techniques, such as Support Vector Machine (SVM), XGBoost, Random Forest, Linear Regression, and Gaussian Naive Bayes, are explored using a comprehensive dataset containing patient demographics, medical history, lifestyle factors, and clinical measurements. Based on different ML models, an Enhanced Stroke Prediction Model (ESPM) is proposed. Grid search, Randomized search, and Bayesian optimization are employed as hyperparameter tuning techniques, and parameters like accuracy, precision, recall, and F1 score are analyzed. It is observed that SVM with Grid Search hyperparameter tunning performs well with an accuracy of 94.129%; Positive Predictive Value (PPV), True Positive Rate(TPR), and F1 Score achieved are 89%, 94%, and 91%, respectively. The outcomes demonstrate the suitability of these models for different aspects of stroke prediction, such as handling complex patterns, capturing non-linearity, robustness to noisy data, and modeling continuous risk scores.
Бесплатно
EUR/USD Exchange Rate Prediction Using Machine Learning
Статья научная
Nowadays artificial intelligence is used in almost every sector of our day-to-day life. AI is used in preventative maintenance, quality control, demand forecasting, rapid prototyping, and inventory management among other places. Also, its use in the economic market has gained widespread. The use of artificial intelligence has made a huge contribution to price forecasting in the currency market or the stock market. This research work explores and analyzes the use of machine learning techniques as a linear regression in the EUR/USD exchange rate in the global forex market to predict future movements and compare daily and hourly data forecasts. As a reason for comparison, linear regression was applied in both hourlies and daily's almost equivalent data sets of the EUR/USD exchange rate and showed differences in results. Which has opened a new door of research on this market. It has been found that the percentage of accuracy of the daily data forecast is higher than the hourly data forecast at the test stage.
Бесплатно
Edge stable sets and secured edge stable sets in hypergraphs
Статья научная
In this paper, we have proved several results regarding edge stable sets and maximal edge stable sets in hypergraphs. We have also proved various results regarding edge stable sets and maximal edge stable sets in partial subhypergraphs. We have introduced the concept of secured edge stable set, maximum secured edge stable set and i_(s )^1- Set in this paper and proved several results about them.
Бесплатно
Effect Neural Networks on Selected Feature by Meta Heuristic Algorithms
Статья научная
Feature selection is one of the issues that have been raised in the discussion of machine learning and statistical identification model. We have provided definitions for feature selection and definitions needed to understand this issue, we check. Then, different methods for this problem were based on the type of product, as well as how to evaluate candidate subsets of features, we classify the following categories. As in previous studies may not have understood that different methods of assessment data into consideration, We propose a new approach for assessing similarity of data to understand the relationship between diversity and stability of the data is selected. After review and meta-heuristic algorithms to implement the algorithm found that the cluster algorithm has better performance compared with other algorithms for feature selection sustained.
Бесплатно
Efficient Optimization of Edge Server Selection Technique in Content Delivery Network
Статья научная
Cloud Computing provides the infrastructure as a "Cloud" from which businesses and users are permit to access applications from anywhere in the world on demand. Thus, the computing world is rapidly transforming towards developing software for millions to consume as a service, rather than to run on their individual computers. But many users could not satisfy on cloud services completely due to their uncovering security purpose for handling large numbers of data. Even the network becomes uncontrollable, when large numbers of user's request to the server create network congestion and data losses vigorously. Content Delivery Network OR CDN is an eminent solution of this problem. Our objective is to create optimized method for edge selection technique in Content Delivery Network to deliver and direct the user request to the nearest edge server and establish the connection between them and transfer the respective content
Бесплатно
Emoji Prediction Using Emerging Machine Learning Classifiers for Text-based Communication
Статья научная
We aim to extract emotional components within statements to identify the emotional state of the writer and assigning emoji related to the emotion. Emojis have become a staple part of everyday text-based communication. It is normal and common to construct an entire response with the sole use of emoji. It comes as no surprise, therefore, that effort is being put into the automatic prediction and selection of emoji appropriate for a text message. Major companies like Apple and Google have made immense strides in this, and have already deployed such systems into production (for example, the Google Gboard). The proposed work is focused on the problem of automatic emoji selection for a given text message using machine learning classification algorithms to categorize the tone of a message which is further segregated through n-gram into one of seven distinct categories. Based on the output of the classifier, select one of the more appropriate emoji from a predefined list using natural language processing (NLP) and sentimental analysis techniques. The corpus is extracted from Twitter. The result is a boring text message made lively after being annotated with appropriate text messages
Бесплатно
Статья научная
Mobile phones and handsets enable us to communicate our voice, data and video messages with individuals that are far-off from us. When an active call is initiated by someone using a mobile phone, it is transmitted through a nearby Base Station (BS) transmitter to another BS until the call gets to its intended receiver. Any time a caller initiates and loses a connection to a BS while on conversation, the call is said to be dropped. The initiation and completion of an active call without any form of disconnection or termination is a key service quality parameter in telecommunication system networks. Robust statistical estimation, modelling and characterization of call drop rates is of high importance to the network operators and radio frequency engineers for effective re-planning and performance management process of telecommunication system networks. This work was designed to determine the optimal probability distribution model for drop call rates based on a five week acquired rate of drop calls data sample in the Southern regions of Nigeria. To accomplish the aim, eight probability distributions namely logistic, log-logistic, normal, log-normal, exponential, Rayleigh, rician and Gumbel max were explored and based on the combined scores of three goodness of fit statistical tests, the log-logistic distribution was found to be the optimal probability distribution for the weekly rate of drop call prognostic analysis. The results could be of immense assistance to radio frequency engineers for optimal statistical modelling and design of cellular systems channels.
Бесплатно
Energy Efficient Multipath Routing in Zone-based Mobile Ad-hoc Networks: Mathematical Formulation
Статья научная
A wireless mobile ad hoc network (MANET) is a dynamic network that can be built without the need for any central governance system and pre-existing infrastructure, in which each node can act as a router. During the transmission of information in any network, energy consumption is an important factor for the efficiency and lifetime of the network. A reduction in energy consumption is achieved by detecting the energy consumption at the node at each stage of transmission. The main objective of this paper is to formulate mathematical models of energy consumption. A mathematical model of the energy consumption of a network is to be built on the basis of available nodes and links to formulate mathematical models related to energy. When constructing this mathematical model for the challenge related to mobility and low connectivity due to limited battery power in the network, the failure of the links present in the network and the estimated energy consumption are taken into account. Due to the greater mobility of this type of network, nodes rapidly change their positions, causing nodes to drain the battery very quickly, thus reducing network performance. So we need a mathematical model which helps in developing a mathematical model after developing a conceptual model. Which helps in predicting the quantitative behavior of a system. Weaknesses and strengths of a model can be identified from the quantitative results of a mathematical model.
Бесплатно
Enhanced E-recruitment using Semantic Retrieval of Modeled Serialized Documents
Статья научная
Retrieval in existing e-recruitment system is on exact match between applicants' stored profiles and inquirer's request. These profiles are captured through online forms whose fields are tailored by recruiters and hence, applicants sometimes do not have privilege to present details of their worth that are not captured by the tailored fields thereby, leading to their disqualification. This paper presents a 3-tier system that models serialized documents of the applicants' worth and they are analyzed using document retrieval and natural language processing techniques for a human-like assessment. Its presentation tier was developed using java server pages and middle tier functionalities using web service technology. The data tier models résumés that have been tokenized and tagged using Brill Algorithm with my sequel. Within the middle tier, indexing was achieved using an inverted index whose terms are noun phrases extracted from résumés that have been tokenized and tagged using Brill Algorithm.
Бесплатно
Evaluating the impact of Test-Driven Development on Software Quality Enhancement
Статья научная
In the software development industry, ensuring software quality holds immense significance due to its direct influence on user satisfaction, system reliability, and overall end-users. Traditionally, the development process involved identifying and rectifying defects after the implementation phase, which could be time-consuming and costly. Determining software development methodologies, with a specific emphasis on Test-Driven Development, aims to evaluate its effectiveness in improving software quality. The study employs a mixed-methods approach, combining quantitative surveys and qualitative interviews to comprehensively investigate the impact of Test-Driven Development on various facets of software quality. The survey findings unveil that Test-Driven Development offers substantial benefits in terms of early defect detection, leading to reduced costs and effort in rectifying issues during the development process. Moreover, Test-Driven Development encourages improved code design and maintainability, fostering the creation of modular and loosely coupled code structures. These results underscore the pivotal role of Test-Driven Development in elevating code quality and maintainability. Comparative analysis with traditional development methodologies highlights Test-Driven Development's effectiveness in enhancing software quality, as rated highly by respondents. Furthermore, it clarifies Test-Driven Development's positive impact on user satisfaction, overall product quality, and code maintainability. Challenges related to Test-Driven Development adoption are identified, such as the initial time investment in writing tests and difficulties adapting to changing requirements. Strategies to mitigate these challenges are proposed, contributing to the practical application of Test-Driven Development. Offers valuable insights into the efficacy of Test-Driven Development in enhancing software quality. It not only highlights the benefits of Test-Driven Development but also provides a framework for addressing challenges and optimizing its utilization. This knowledge is invaluable for software development teams, project managers, and quality assurance professionals, facilitating informed decisions regarding adopting and implementing Test-Driven Development as a quality assurance technique in software development.
Бесплатно
Evidential paradigm and SAD systems: features and peculiarities
Статья научная
Research on automated reasoning systems based on a number of paradigms that support human activity in formalized text processing began in the late 1950s – early 1960s, when computer performance and memory space became sufficient for programming of complex intelligent processes. The so-called evidential paradigm was among them and it can be viewed as a way for integrating all reasonable paradigms oriented to the development of computer languages for representing formalized texts in the form most suitable for a user, formalization and development of the evidence of a computer-made proof step, creation of the information environment having influence on a current evidence of a machine proof step, and an active human-machine interaction. This work contains a brief description of the evidential paradigm and its implementation in the form of intelligent systems intended for the symbolic and deductive processing of mathematical texts focusing main attention on their features and peculiarities.
Бесплатно
Exact Analytical Solution of Boundary Value Problem in a Form of an Infinite Hypergeometric Series
Статья научная
This paper proposes an exact solution of the classical Graetz problem in terms of an infinite series represented by a nonlinear partial differential equation considering two space variables, two boundary conditions and one initial condition. The mathematical derivation is based on the method of separation of variables whose several stages were illustrated to reach the solution of the Graetz problem.
Бесплатно
Ferrer diagram based partitioning technique to decision tree using genetic algorithm
Статья научная
Decision tree is a known classification technique in machine learning. It is easy to understand and interpret and widely used in known real world applications. Decision tree (DT) faces several challenges such as class imbalance, overfitting and curse of dimensionality. Current study addresses curse of dimensionality problem using partitioning technique. It uses partitioning technique, where features are divided into multiple sets and assigned into each block based on mutual exclusive property. It uses Genetic algorithm to select the features and assign the features into each block based on the ferrer diagram to build multiple CART decision tree. Majority voting technique used to combine the predicted class from the each classifier and produce the major class as output. The novelty of the method is evaluated with 4 datasets from UCI repository and shows approximately 9%, 3% and 5% improvement as compared with CART, Bagging and Adaboost techniques.
Бесплатно
Forecasting Natural Gas Prices Using Nonlinear Autoregressive Neural Network
Статья научная
When forecasting time series, It was found that simple linear time series models usually leave facets of economic and financial unknown in the forecasting time series due to linearity behavior, which remains the focus of empirical and applied study. The study suggested the Nonlinear Autoregressive Neural Network model and a comparison was made using the ARIMA model for forecasting natural gas prices, as obtained from the analysis, NAR models were better than the completed ARIMA model, measured against three performance indicators. The decision criterion for the selection of the best suited model depends on MSE, RMSE and R2. From the results of the criterion it has found that both the models are providing almost closed results but NAR is the best suited model for the forecasting of natural gas prices.
Бесплатно
Future Possible Age of the Universe with Density Variation
Статья научная
A fundamental principle and assumption of cosmology says that the universe is homogeneous and isotropic when viewed on a large scale. According to the cosmological principle, space might be flat, or have a negative or positive curvature in cosmological model. Positively curved universe denotes the closed universe and negatively curved universe denotes the open universe. Our universe type is flat because it expands in every direction neither curving positively nor negatively. We have observed that the progression of the universe is based on radiation and matter domination. In this paper we also have observed that future possible upper limit age of the universe is 9.4203×〖10〗^10 years which varies with density.
Бесплатно
Статья научная
Regression testing is used to check that changes in the some functionality of the software to not affect its old behaviours. Test case prioritization is essential for reducing the cost of regression testing. In this paper a test cases prioritization model based on fuzzy logic is presented. State machine diagram is used to capture the behaviour of the system. Risk information is associated with the states. After change in the functionality of the system new state machine diagram is designed. This new state machine diagram is converted into Weighted Extended Finite State Machine (WEFSM). Weights are assigned to nodes and edges based on change and risk exposure. Risk exposure and change information of each test case is used as input to fuzzy model. Test cases are categorized in retestable, reusable and obsolete.
Бесплатно
Generalized Algorithm on idiosyncrasy of Numbers
Статья научная
In this paper, using computer algebra system a new generalized algorithm is developed to study and generalize the Kaprekar’s operation which can be used for desired numbers of iterations and is also applicable to any n-digits number which is greater than or equal to two. Existing relevant results are verified with the available results in literature and further extended to examine the difference (kernel) of the obtained number during the process with the number obtained in preceding iteration after each step. Sum of the digits of an acquired number obtained after each step is also noticed and found that sum of its digits is divisible by 9. A detailed investigation is conducted for all two-digit number and the output acquired is exhibited in tabular form which has not been studied in earlier. An 8-digits number also considered and found that it does not converges to a unique kernel like 3-digits and 4-digits, but follows a regular pattern after initial iteration. Analytical illustrations are provided along with pictorial representations for 2-digits, 3-digits 4-digits and 8-digits number. This algorithm can further be employed for numbers having any number of digits.
Бесплатно
Graph Dynamic Threshold Model Resource Network: Key Features
Статья научная
In this paper, we describe a graph dynamic threshold model called resource network, and briefly present the main results obtained during several years of research. Resource Network is represented by a connected oriented with weighted graph with an arbitrary topology. Weights of edges denote their throughput capacities for an abstract resource. The resource is stored in vertices, which can contain its unlimited amount. Network operates in discrete time. The total amount of resource is constant, while pieces of resource are reallocating among vertices every time step, according to certain rules with threshold switching. The main objective of our research is to define for a network with an arbitrary topology all its basic characteristics: the vectors of limit state and flow for every total amount of resource W; the threshold value of total recourse T, which switches laws of operating of the network; description of these laws. It turned out that there exists several classes of networks depending on their topologies and capacities. Each class demonstrates fundamentally different behavior. All these classes and their characteristics will be reviewed below.
Бесплатно
Green Computing: An Era of Energy Saving Computing of Cloud Resources
Статья научная
Cloud computing is a widely acceptable computing environment, and its services are also widely available. But the consumption of energy is one of the major issues of cloud computing as a green computing. Because many electronic resources like processing devices, storage devices in both client and server site and network computing devices like switches, routers are the main elements of energy consumption in cloud and during computation power are also required to cool the IT load in cloud computing. So due to the high consumption, cloud resources define the high energy cost during the service activities of cloud computing and contribute more carbon emissions to the atmosphere. These two issues inspired the cloud companies to develop such renewable cloud sustainability regulations to control the energy cost and the rate of CO2 emission. The main purpose of this paper is to develop a green computing environment through saving the energy of cloud resources using the specific approach of identifying the requirement of computing resources during the computation of cloud services. Only required computing resources remain ON (working state), and the rest become OFF (sleep/hibernate state) to reduce the energy uses in the cloud data centers. This approach will be more efficient than other available approaches based on cloud service scheduling or migration and virtualization of services in the cloud network. It reduces the cloud data center's energy usages by applying a power management scheme (ON/OFF) on computing resources. The proposed approach helps to convert the cloud computing in green computing through identifying an appropriate number of cloud computing resources like processing nodes, servers, disks and switches/routers during any service computation on cloud to handle the energy-saving or environmental impact.
Бесплатно
Handling Numerical Missing Values Via Rough Sets
Статья научная
Many existing industrial and research data sets contain missing values. Data sets contain missing values due to various reasons, such as manual data entry procedures, equipment errors, and incorrect measurements. It is usual to find missing data in most of the information sources used. Missing values usually appear as "NULL" values in the database or as empty cells in the spreadsheet table. Multiple ways have been used to deal with the problem of missing data. The proposed model presents rough set theory as a technique to deal with missing data. This model can handle the missing values for condition and decision attributes, the web application was developed to predict these values.
Бесплатно
