Cloud Security Reinvented: A Predictive Algorithm for User Behavior-Based Threat Scoring
Uložené v:
| Názov: | Cloud Security Reinvented: A Predictive Algorithm for User Behavior-Based Threat Scoring |
|---|---|
| Autori: | Sudhakara Reddy Peram |
| Zdroj: | Journal of Business Intelligence and Data Analytics. 2:1-9 |
| Informácie o vydavateľovi: | Sciforce LLC, 2025. |
| Rok vydania: | 2025 |
| Popis: | This study presents a robust algorithmic approach to evaluating login security behavior using multi-criteria analysis. By integrating parameters such as login attempts, session duration, and data upload volumes, the study aims to quantify user activity risks and enhance security threat detection. The developed model calculates a Threat Risk Score to evaluate potential threats across diverse user profiles. The proposed methodology facilitates proactive identification of abnormal behaviors, which is critical for real-time cybersecurity operations. Research Significance: In an era where cybersecurity threats are increasingly sophisticated, identifying risky user behaviors through data-driven analysis is of paramount importance. This research contributes significantly by offering a novel threat evaluation framework based on behavioral parameters. The approach allows organizations to detect potential security breaches early, thereby reducing the attack surface and improving response efficiency. Methodology: The methodology is centered on the design and implementation of an intelligent evaluation algorithm that incorporates three behavioral attributes: Login_Attempts, Avg_Session_Duration_Min, and Data_Upload_MB. These alternatives are normalized and analyzed using a weighted decision-making algorithm to derive a composite The model integrates threshold analysis and pattern recognition to ensure accurate threat classification and anomaly detection. Alternative: The alternatives evaluated in this study are derived from user session data: Login Attempts: Frequency of user login trials within a defined time window. Avg Session Duration Min: The average duration of each user session, representing usage intensity. Data Upload MB: The total volume of data uploaded during the session, indicating potential data exfiltration. These features are selected based on their strong correlation with known threat patterns. Evaluation Parameter: Threat Risk Score is used as the principal evaluation metric. It is computed by aggregating normalized values of the three behavioral alternatives, adjusted using pre-defined risk weightings. A higher score signifies a greater probability of anomalous or malicious behavior, enabling swift prioritization for security response teams. Result: The algorithm was tested on a synthetic dataset simulating diverse user behaviors. Results show high accuracy in distinguishing between normal and high-risk activities, with an overall detection precision exceeding 90%. The model effectively prioritizes threats based on behavioral deviations and demonstrates its applicability for real-world security monitoring systems. |
| Druh dokumentu: | Article |
| ISSN: | 2998-3541 |
| DOI: | 10.55124/jbid.v2i3.252 |
| Prístupové číslo: | edsair.doi...........a4712583f51374e47a8c0d323ebb443d |
| Databáza: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstrakt: | This study presents a robust algorithmic approach to evaluating login security behavior using multi-criteria analysis. By integrating parameters such as login attempts, session duration, and data upload volumes, the study aims to quantify user activity risks and enhance security threat detection. The developed model calculates a Threat Risk Score to evaluate potential threats across diverse user profiles. The proposed methodology facilitates proactive identification of abnormal behaviors, which is critical for real-time cybersecurity operations. Research Significance: In an era where cybersecurity threats are increasingly sophisticated, identifying risky user behaviors through data-driven analysis is of paramount importance. This research contributes significantly by offering a novel threat evaluation framework based on behavioral parameters. The approach allows organizations to detect potential security breaches early, thereby reducing the attack surface and improving response efficiency. Methodology: The methodology is centered on the design and implementation of an intelligent evaluation algorithm that incorporates three behavioral attributes: Login_Attempts, Avg_Session_Duration_Min, and Data_Upload_MB. These alternatives are normalized and analyzed using a weighted decision-making algorithm to derive a composite The model integrates threshold analysis and pattern recognition to ensure accurate threat classification and anomaly detection. Alternative: The alternatives evaluated in this study are derived from user session data: Login Attempts: Frequency of user login trials within a defined time window. Avg Session Duration Min: The average duration of each user session, representing usage intensity. Data Upload MB: The total volume of data uploaded during the session, indicating potential data exfiltration. These features are selected based on their strong correlation with known threat patterns. Evaluation Parameter: Threat Risk Score is used as the principal evaluation metric. It is computed by aggregating normalized values of the three behavioral alternatives, adjusted using pre-defined risk weightings. A higher score signifies a greater probability of anomalous or malicious behavior, enabling swift prioritization for security response teams. Result: The algorithm was tested on a synthetic dataset simulating diverse user behaviors. Results show high accuracy in distinguishing between normal and high-risk activities, with an overall detection precision exceeding 90%. The model effectively prioritizes threats based on behavioral deviations and demonstrates its applicability for real-world security monitoring systems. |
|---|---|
| ISSN: | 29983541 |
| DOI: | 10.55124/jbid.v2i3.252 |