RideScan

Guidance

Connection Failures

We've tried to create the service as robust as possible, but sometimes connection failures can occur due to network issues, server downtime, or incorrect configurations. We've listed down a few steps you can take to troubleshoot and resolve connection failures:

Upload Interruptions

There are situations where the upload process may be interrupted due to various reasons such as network instability, file size limitations, or server issues which might affect the next step of calibration of risk assessment. We encourage you to follow these steps to troubleshoot and resolve upload interruptions:

  • Check your internet connection and ensure it is stable.
  • Verify that the file size does not exceed the maximum limit.
  • Ensure that the server is up and running without any issues.

Once the issue has been resolved, you can retry the process. We recommend you to retrying the calibration process afterwards once again if the problem occured during calibration.
As for the risk assessment, you would only need to upload the files that haven't been uploaded and processed earlier yet.

Authentication / permissions errors

Almost all request require you to be authenticated and have the necessary permissions to access the resources. If you encounter authentication or permissions errors, please follow these steps:

  • Check your API key and ensure it is valid. You can go to the API settings page to regenerate it if needed.
  • Verify that you have the necessary permissions to access the resource.
  • Ensure that your user account is in good standing and not suspended.

Once the issue has been resolved, you can retry the process. We recommend you to retrying the calibration process afterwards once again if the problem occured during calibration.
As for the risk assessment, you would only need to upload the files that haven't been uploaded and processed earlier yet.

Server unavailable / timeout

If you encounter server unavailable or timeout errors, please follow these steps:

  • Check the server status page to see if there are any ongoing issues.
  • Try again later, as the issue may be temporary.
  • If the problem persists, contact support for assistance.

Best practices

To ensure smooth uploads and processing, follow these best practices:

  • Ensure the file integrity before uploading them.
  • Keep your API key secure and do not expose it in client-side code.
  • Use descriptive names for your files and missions to make them easily identifiable.
  • Regularly monitor the status of your uploads and processing jobs.
  • Implement error handling and retries for failed requests.

Misclassified Anomalies

Misclassified anomalies are instances where an anomaly is incorrectly labeled or categorized. This can happen due to various reasons, such as insufficient training data, model bias, or changes in the underlying data distribution. Misclassified anomalies can lead to incorrect conclusions and actions.

Improving accuracy with more data

To improve the accuracy of your anomaly detection system, consider the following strategies:

  • Collect more diverse and representative data during the calibration process.
  • Augment your existing dataset with synthetic data to cover edge cases.
  • You can always re-calibrate the model to keep it relevant with more updated routine data.

Understanding risk threshold

To effectively manage risks in your anomaly detection system, consider the following strategies:

  • The score is represented as a probability between 0 and 100, indicating the likelihood of an anomaly.
  • Establish a threshold score that determines when an anomaly should be investigated further.
  • A score above the critical threshold may require immediate attention and action.