Time Series Forecasting Using Foundation Models
Деталі
Ціни
Відгуки
Time Series Forecasting Using Foundation Models
Ще ніхто не оцінив.
Про книжку
Your forecasts lag while data grows, hardware costs, and deadlines tighten. Traditional model training demands weeks of tuning and GPU burns time. Meanwhile, foundation models already understand seasonality, holiday spikes, and rare shocks. This book hands you TimeGPT, Chronos, and other pretrained powerhouses. Generate zero-shot forecasts or fine-tune quickly with only laptop resources. Deliver stronger predictions, faster insights, and measurable business value in days, not months.
Model internals explained: Understand how large time models capture temporal patterns and uncertainty.
Zero-shot workflow: Run instant forecasts on custom data without retraining, saving weeks of effort.
Fine-tuning guides: Adapt foundation models to niche domains for even higher accuracy.
Evaluation playbook: Benchmark probabilistic and point forecasts using industry-standard metrics.
Laptop-friendly code: All examples rely on Python and CPUs, no high-end GPUs required.
Time Series Forecasting Using Foundation Models by data-science instructor Marco Peixeiro containing clear diagrams, annotated notebooks, and rigorously tested examples establish immediate credibility.
You build a tiny foundation model to grasp pretraining mechanics, then experiment with production-grade models like TimeGPT and Chronos. Each chapter layers hands-on labs, checkpoints, and real-world case studies. Finish ready to integrate pretrained forecasting models, slash development time, and present trustworthy predictions to stakeholders. Your pipeline becomes faster, cheaper, and easier to maintain.
Designed for data scientists and ML engineers comfortable with basic forecasting theory and Python.
Деталі
- Видання
- Manning, 2025, паперова
- Видавництво
- Manning
- Сторінок
- 256
- Мова
- Англійська
- ISBN
- 9781633435896
Відгуки
Відгуків ще немає.
Ціни 1 пропозиція
Деякі посилання на магазини — партнерські, ми можемо отримати невеликий відсоток від магазину, але це не впливає на порядок пропозицій на сайті.
