USE OF MACHINE-LEARNING TECHNIQUES BASED ON PYTHON LANGUAGE CODE TO CLASSIFY FAILURE DATA FROM THE BRAZILIAN AIR FORCE DATABASE

Aug 5, 2026

Upload by Eliana

The Brazilian Air Force (BrAF) must manage multiple aircraft fleets and rely on accurate failure data to support decisions on reliability, availability, and maintainability. BrAF engineers currently preprocess records by repeatedly classifying entries as failure or non-failure, a process that is labor intensive and time consuming. This study develops and evaluates a machine-learning model to automate that classification task. Six machine-learning techniques are assessed, and the Support Vector Classifier (SVC) delivers the best results using the F1-score metric. The findings indicate that SVC can classify failure data from the BrAF database accurately while reducing effort, improving consistency in failure recording, and preventing non-useful database entries for better inventory-related repair information.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages153

Reading Time6 h 26 min

AccessFree to download

Abstract

Problem and Motivation

Proposed Machine-Learning Approach

Model Evaluation and Results

Implications for Failure Recording and Inventory Management

Calhoun: The NPS Institutional Archive DSpace Repository | 2023-06
USE OF MACHINE-LEARNING TECHNIQUES BASED ON PYTHON LANGUAGE CODE TO CLASSIFY FAILURE DATA FROM THE BRAZILIAN AIR FORCE DATABASE
de Almeida, Ygor H.
Monterey, CA; Naval Postgraduate School | | --- | | [https://hdl.handle.net/10945/72156](https://hdl.handle.net/10945/72156) | NPS Scholarship Theses Copyright is reserved by the copyright owner. Downloaded from NPS Archive: Calhoun NAVAL POSTGRADUATE SCHOOL MONTEREY, CALIFORNIA THESIS USE OF MACHINE-LEARNING TECHNIQUES BASED ON PYTHON LANGUAGE CODE TO CLASSIFY FAILURE DATA FROM THE BRAZILIAN AIR FORCE DATABASE by Ygor H. de Almeida June 2023 Thesis Advisor: Susan K. Aros Co-Advisor: Daniel Cherobini, Brazilian Air Force Approved for public release. Distribution is unlimited. THIS PAGE INTENTIONALLY LEFT BLANK | REPORT DOCUMENTATION PAGE | | | | | Form Approved OMB No. 0704-0188 | | | --- | --- | --- | --- | --- | --- | --- | | Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instruction, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704- 0188) Washington, DC, 20503. | | | | | | | | 1. AGENCY USE ONLY (Leave blank) | | 2. REPORT DATE
June 2023 | | 3. REPORT TYPE AND DATES COVERED
Master’s thesis | | | | 4. TITLE AND SUBTITLE
USE OF MACHINE-LEARNING TECHNIQUES BASED ON PYTHON LANGUAGE CODE TO CLASSIFY FAILURE DATA FROM THE
BRAZILIAN AIR FORCE DATABASE | | | | | 5. FUNDING NUMBERS | | | 6. AUTHOR(S) Ygor H. de Almeida | | | | | | | | 7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)
Naval Postgraduate School
Monterey, CA 93943-5000 | | | | | 8. PERFORMING
ORGANIZATION REPORT NUMBER | | | 9. SPONSORING / MONITORING AGENCY NAME(S) AND
ADDRESS(ES)
N/A | | | | | 10. SPONSORING / MONITORING AGENCY REPORT NUMBER | | | 11. SUPPLEMENTARY NOTES The views expressed in this thesis are those of the author and do not reflect the official policy or position of the Department of Defense or the U.S. Government. | | | | | | | | 12a. DISTRIBUTION / AVAILABILITY STATEMENT
Approved for public release. Distribution is unlimited. | | | | | 12b. DISTRIBUTION CODE
A | | | 13. ABSTRACT (maximum 200 words)
Like other major flight operators, the Brazilian Air Force (BrAF) must effectively manage multiple aircraft fleets, which are costly assets. The failure data collected from these assets is essential for decisionmakers to assess the systems’ reliability, availability, and maintainability. Obtaining accurate and reliable information depends on the quality of failure data collected. BrAF engineers typically preprocess the data by classifying it as failure or non-failure for analysis, but this task is repetitive and time-consuming. Therefore, this study aims to develop and evaluate a machine-learning model capable of automatically performing this classification task. Of the six machine-learning techniques assessed, the Support Vector Classifier (SVC) model performed best in the F1-score metric. The results suggest that the SVC model has the potential to classify failure data from the BrAF database accurately, saving a significant amount of time. Additionally, the model could aid maintainers during the failure recording process, preventing them from inserting non-useful data in the database, and for inventory management of specific workshop repairs, thus providing more acc

FAQs

What classification task does the study focus on?

The study automates classifying failure data into failure versus non-failure labels for analysis.

Why is accurate failure data important for the Brazilian Air Force?

It is essential for decision-makers to assess systems’ reliability, availability, and maintainability.

Which machine-learning technique performed best and by what metric?

The Support Vector Classifier (SVC) performed best, evaluated using the F1-score metric.

Aug 5, 2026

Upload by Eliana