This is an old revision of the document!
Forecasting Meeting Minutes for Week of September 27, 2016
Present: Gordon, Brieanna, Jaimie
Updates
- More coordination with Seyyed
Expectations
- 1. Well documented code
- Inputs/Outputs:
- Mean Squared Error
- Input
- Weight of length
- 2. Learning signal processing/machine learning algorithms
- Reproducing what Masaki did
- Visualizing Data w/ 3D plotting:
- Contact Masaki
- Work on normalized data with Zenith Angle
- 3. Forecast Solar Irradiance
- Recursive least squares
- Least mean squares
- Machine learning aspect
- 4. Understanding solar data
- PV and solar data (non-linear relationship)
- Weather effects sudden changes in weather
- Reserve or storage energy
- Shift demands
Machine Learning
Know linear, unsupervised and classification * 2 types of algorithms:
- Regression/Estimation
- Classification (binary/descrete)/Detection
- Classification: Iris Problem
- Linear threshold function: Hyperplane separating data (0 | 1)
- 2 items from 1 item = possible
* Unsupervised Learning:
- Cluster
- PCA
* Support Vector Machines:
- Non-linear
* Multi-learning networks:
- Deep learning
* Andrew Ing coursera