| --- |
| language: |
| - en |
| license: other |
| license_name: bsl-1.1 |
| license_link: LICENSE |
| tags: |
| - time-series |
| - time-series-forecasting |
| - forecasting |
| - gift-eval |
| - meta-learning |
| - agentic |
| - pytorch |
| datasets: |
| - Salesforce/GiftEval |
| pipeline_tag: time-series-forecasting |
| --- |
| |
| # CompactTS-Agent: Autonomous Agentic Meta-Forecaster for Time Series |
|
|
| Developed by **Daksh Bhalala**. |
|
|
| `CompactTS-Agent` is an autonomous agentic meta-forecasting system designed for general time-series forecasting across diverse industrial domains (energy grids, server telemetry, traffic flow, nature observations, and healthcare). |
|
|
| ## Architecture Overview |
|
|
| Rather than relying on a single static neural network across disparate domains, `CompactTS-Agent` operates as an autonomous orchestration system: |
|
|
| 1. **Scale-Invariant Series Profiler:** Extracts structural statistical and spectral signatures from incoming series (Fourier spectral entropy $H_s$, normalized mean drift $\Delta_\mu$, and ARCH volatility clustering $R^2_{\text{ARCH}}$). |
| 2. **Meta-Routing Policy:** Deterministically maps time-series physical regimes to specialized foundation model representations (e.g., concentrated harmonics, low-drift telemetry, volatility clustering). |
| 3. **Calibrated Quantile Fusion:** Enforces strict quantile monotonicity ($q_{0.1} \le q_{0.2} \le \dots \le q_{0.9}$) and physical support projections. |
|
|
| ## Installation & Quickstart |
|
|
| ```bash |
| git clone https://github.com/DakshBhalala/CompactTS-Agent.git |
| cd CompactTS-Agent |
| pip install -e . |
| ``` |
|
|
| ### Python API Usage |
|
|
| ```python |
| import numpy as np |
| import pandas as pd |
| from compactts.agent import CompactTSAgent, SeriesProfiler |
| |
| # 1. Profile an incoming time series |
| profiler = SeriesProfiler(freq="H") |
| series = np.sin(np.linspace(0, 20, 100)) + 0.1 * np.random.randn(100) |
| profile = profiler.profile(series) |
| print("Series Profile:", profile) |
| |
| # 2. Generate calibrated probabilistic forecasts |
| agent = CompactTSAgent(prediction_length=24, freq="H") |
| start_dates = [pd.Period("2024-01-01 00:00", freq="h")] |
| forecasts = agent.forecast_batch([series], start_dates) |
| |
| # Access forecast quantiles (9 levels: 0.1 to 0.9) |
| q_forecasts = forecasts[0].forecast_array |
| print("Forecast shape (9 quantiles, 24 horizon):", q_forecasts.shape) |
| ``` |
|
|
| ## Benchmark Evaluation |
|
|
| CompactTS-Agent evaluation results are submitted under `model_type: "agentic"`. |
|
|
| To reproduce results on official benchmark configurations: |
| ```bash |
| python scripts/run_agentic_benchmark.py |
| ``` |
|
|
| ## Commercial Licensing & Enterprise Deployment |
|
|
| CompactTS-Agent is licensed under the **Business Source License 1.1 (BSL 1.1)**: |
| - **Free for Non-Commercial Research & Benchmark Replication:** Anyone is free to inspect, run, evaluate, test, and reproduce benchmarks using CompactTS-Agent. |
| - **Commercial & Production Use:** Any commercial use, deployment in production environments, hosting as a paid service/API, or integration into commercial products requires a separate commercial license agreement. |
|
|
| For commercial inquiries, enterprise deployment rights, or proprietary licensing: |
| - **Author:** Daksh Bhalala |
| - **Code Repository:** [DakshBhalala/CompactTS-Agent](https://github.com/DakshBhalala/CompactTS-Agent) |
| - **License:** Business Source License 1.1 (BSL-1.1) |
|
|
