Cloud task failure prediction Algorithm

S1. Get normalized dataset for cloud task fault prediction.

S2. Build feature set for cloud task fault prediction.

S3. Create initial multi-scale GRU model.

S4. Initialize Sparrow Search Algorithm population matrix.

S5. Output optimal hyperparameters for cloud task fault prediction.

S6. Get trained model parameters for cloud task fault prediction.

S7. Deploy trained model to cloud platform, generate fault probability sequence, compare against risk threshold, and output early warning when threshold exceeded.

S8. Send fault warning to cloud platform scheduling module.

This invention reduces resource waste and service interruption risk from sudden task failures.