sikaiyin / load-forcasting-with-distributed-dl

Sichuan University Undergraduate Thesis Research: The short-term load forecasting based on big data processing algorithms

load-forcasting-with-DL

The undergraduate thesis on load forcasting model built on spark and tensorflow. There are mainly two parts of this research work:

1. Load forecasting via LSTM Neural Network

We built a 2-layer LSTM time-series forecasting nerual network for the power load data.

2. Distributed ML models based on TensorFlowOnSpark Architecture

We built a distributed deep learning system by combining the Apache Spark, Apache Hadoop and Distributed TF. The system is based on the architecture of TensorFlowOpSpark by Yahoo.