Purpose <p>Since the advent of the Industry 4.0 era, the demand for intelligent industrial machining that emphasizes high precision and optimized operational costs has risen significantly. Signal processing methodologies offer promising solutions to enhance condition monitoring challenges. However, achieving more accurate methods and effective data volume management remains an open problem.</p> Methods <p>One promising method is the Reduced-order Symptom Recognition (RSR) technique, which helps track the condition of tools by analyzing time-frequency spectrums. This is particularly useful for studying signals that change over time, like the vibrations from bearings that are wearing out. However, the RSR method still has some areas that need improvement, such as accuracy in both time and frequency analysis and its ability to manage data volume effectively. To address these issues, we have developed an Enhanced Reduced-order Symptom Recognition (ERSR) method. This approach uses Pearson's correlation coefficient to make the time-frequency representations clearer.</p> Results <p>We tested the ERSR method using the NASA-FEMTO bearing dataset, which includes signals from normal operation to failure. By comparing our method with other established time-frequency techniques, we have showed that the ERSR method not only provides accurate spectra but also helps control the amount of data effectively. </p> Conclusion <p>This research introduces the Enhanced Reduced-order Symptom Recognition (ERSR) method, which aims to provide clearer joint time-frequency spectrums while minimizing data volume without sacrificing accuracy. The effectiveness of this method has been confirmed through comparisons with established techniques, making it useful for practical condition monitoring applications, particularly for diagnosing bearing faults.</p>