Reproducible Benchmark · NAB + ETT Datasets

33,000–48,000 Predictions/Second. Lower Error Than Deep Learning.

DriftMind was benchmarked against Adaptive ARIMA, Prophet, and OneNet — the leading deep learning architecture for online forecasting. It ran on a standard CPU with no GPU, no pre-training, and no retraining.

Results Summary

Metric Value
Predictions per second (CPU) 33–48K
MAE on NAB Machine Temp. 0.82
Speed compared to OneNet 140× faster (25 seconds vs 58 minutes)
Retraining cycles 0 (vs up to 17,486 for Prophet)

Benchmark 1 — NAB Multi-Dataset

Three-Way Direct Comparison

All three models were evaluated on six datasets from the Numenta Anomaly Benchmark (NAB) — spanning machine sensors, cloud infrastructure, urban demand, ad-exchange pricing, and social-media volume — at three forecast horizons (h = 1, 6, 24). Results are reported per horizon:

Model MAE (lower is better) Throughput (pred/s) Total Time Retraining Cycles
DriftMind 0.8526 10,274 2.2 s Continuous
Adaptive ARIMA 0.8529 47 8m 3s 901
Triggered Prophet 2.9901 13 29m 0s 4,105

Key Finding

On seasonal workloads (NYC Taxi, CPU Utilization, Twitter Volume), DriftMind wins accuracy at every horizon — by 1.4–5× at h = 6 and h = 24 — and is orders of magnitude faster. On non-seasonal, low-amplitude signals (Machine / Ambient Temperature, Exchange-2), ARIMA is competitive or better at short and long horizons.

Benchmark 2 — ETTh2 & ETTm1 vs OneNet

DriftMind vs. State-of-the-Art Deep Learning

DriftMind was benchmarked against OneNet, the leading deep learning architecture for online time series forecasting under concept drift.

Model MAE MSE Total Runtime Hardware Warm-up Required
DriftMind 0.232 0.145 00:00:25 CPU only None (cold-start)
OneNet (NeurIPS 2023) 0.348 0.380 00:58:32 RTX 3080 Ti GPU 25% dataset

What this shows

On ETTh2, DriftMind is both faster and more accurate than OneNet. On ETTm1, OneNet achieves a marginal accuracy edge at longer horizons, while DriftMind remains 27–28× faster and requires no warm-up window.

The Latency of Learning

Retraining Lag

ARIMA and Prophet rebuild internal models as new data arrives, leading to structural lag in production systems.

Continuous Adaptation

DriftMind does not pause, does not retrain, and does not reprocess history. Each new observation updates the internal representation immediately, sustaining thousands of predictions per second.

CPU-Only Economics

Running on commodity hardware eliminates GPU cluster costs, making DriftMind economically advantageous in environments with many time series.

Benchmark Methodology

Setup & Reproducibility

Both benchmarks were designed to expose all models to the same data in the same streaming order without cherry-picking or hyperparameter tuning.

1. NAB Dataset

Machine Temperature System Failure series with 22,695 data points.

2. ETTh2 & ETTm1 Datasets

Publicly available datasets for fair comparison.

3. Hardware

DriftMind: Intel Core i7-12700K @ 3.60GHz, 32 GB DDR4. OneNet: same machine, NVIDIA RTX 3080 Ti.

4. Cold-Start Condition

DriftMind begins predictions from the first observation. OneNet requires 25% of the dataset as pre-training.

5. OneNet Replication

The official OneNet GitHub implementation was used, confirming benchmark integrity.

6. DriftMind Settings

Input length: 20–60, Max clusters: 200, Sliding window gap rate: 2.0.

Reproduce the NAB benchmark yourself — run the full comparison locally with a single Docker command.

Run DriftMind on Your Own Data

Results in seconds.