machine learning

Optimizing Road Traffic Through Reinforcement Learning

In the article, modern approaches to the development of Intelligent Transportation Systems (ITS) aimed at optimizing urban traffic are analyzed. Special attention is paid to model-free reinforcement learning algorithms (Q-Learning and Deep Q-Learning) used for controlling traffic lights in dynamic road traffic conditions. Simulation results in the SUMO environment have proven that implementing such algorithms significantly reduces intersection queues and increases the capacity of the transportation network.

Digital Tools in the Energy Drink Market

In the current context of digital transformation within the energy drinks market, the use of digital technologies has become a crucial tool for enhancing the efficiency of business processes, marketing strat- egies, and consumer engagement. However, despite considerable opportunities, the widespread imple- mentation of digital instruments in this sector faces several challenges that require both academic analysis and practical solutions. One of the key issues is the adaptation of energy drink producers' business models to the realities of the digital environment.

OPTIMIZATION OF THE ROUTING PROCESS IN DISTRIBUTED NETWORKS USING MACHINE LEARNING

The article proposes an innovative approach to optimize the routing process in distributed networks using machine learning techniques, specifically reinforcement learning. This method enables the adaptive determination of optimal data transmission paths based on current network conditions, enhancing overall performance and resilience to dynamic traffic fluctuations. The proposed approach dynamically adjusts to variations in network topology, traffic load, and node availability, ensuring efficient data flow management even in highly dynamic environments.

APPLICATION OF MACHINE LEARNING FOR USER SENTIMENT ANALYSIS IN INFORMATION AND COMMUNICATION SYSTEMS

The article examines modern methods of applying machine learning and recommendation systems for sentiment analysis of users in information and communication environments. Social networks and digital platforms have become important sources of public opinion, generating large volumes of textual data daily. Traditional analysis methods, such as lexical approaches or classical machine learning algorithms, have limitations in detecting context, sarcasm, slang, and emotional nuances in the text. This complicates the accurate identification of user emotions and socially significant topics.

SHAP-BASED EVALUATION OF FEATURE IMPORTANCE IN BGP ANOMALY DETECTION MODELS

The classification of Border Gateway Protocol (BGP) anomalies is essential for maintaining Internet stability and security, as such anomalies can impair network functionality and reliability. Previous studies has examined the impact of key features on anomaly detection; however, current methodologies frequently demonstrate high computational costs, complexity, and usage challenges.

HYBRID MODEL OF NETWORK ANOMALIES DETECTION USING MACHINE LEARNING

The increasing complexity of cyber threats requires the development of effective methods for detecting and classifying attacks in network traffic. This study analyzes the effectiveness of three popular machine learning algorithms: Random Forest, which is used for anomaly detection, Support Vector Machines (SVM), which performs cyber threat classification, and autoencoders, which are used for data preprocessing and deep traffic analysis.

MODELING OF ARTIFICIAL INTELLIGENCE FOR REVENUE FORECASTING

In this study, it was analyzed the ability to forecast the revenues of major corporations such as Apple, Amazon, GE, IBM, and ExxonMobil using Random Forest and XGBoost machine learning algorithms, as well as Tableau as a benchmark analytics tool. The main objective was to assess the accuracy of these methods and their capability to predict financial indicators based on historical data. Google Colab was used as the computational environment, which enabled seamless integration of algorithms, handling of large datasets, and rapid model testing.

Hybrid least squares support vector machine for water level forecasting

Previous studies have highlighted the significant role of historical water level data in flood forecasting.  In this study, we compare two standalone models, Support Vector Machine (SVM) and Least Squares Support Vector Machine (LSSVM), with hybrid models that integrate Ensemble Empirical Mode Decomposition (EEMD) with SVM and LSSVM, aiming to develop a more effective forecasting approach for hydrological data.  Particle Swarm Optimization (PSO) is incorporated into these hybrid models to optimize the parameters of SVM and LSSVM, resulting in four models: SVM-PSO, LSSVM

MODELS FOR TIME SERIES FORECASTING USING ARIMA AND LSTM IN ECONOMICS AND FINANCE

Time series forecasting is a crucial task in economics, business, and finance. Traditionally, forecasting methods such as autoregression (AR), moving average (MA), exponential smoothing (SES), and, most commonly, the autoregressive integrated moving average (ARIMA) model are used. The ARIMA model has demonstrated high accuracy in predicting future time series values. With the advancement of computational power and deep learning algorithms, new approaches to forecasting have emerged.

METHODOLOGY FOR IMPLEMENTING SELF-LEARNING FEEDBACK MODELS IN CRM SYSTEMS: COMPARATIVE ANALYSIS OF EFFECTIVENESS

The article proposes a methodology for implementing self-learning feedback models in Customer Relationship Management (CRM) systems. It examines the key issues of existing CRM systems, including insufficient adaptability to changes in customer behavior and limited capabilities for automatic data analysis.