https://padlet.com/musfirahdamla/ANNweek7 Jangan tidur en-us 2017-03-28 06:31:45 UTC 2026-02-09 19:25:55 UTC hello@padlet.com https://padlet-assets.s3.amazonaws.com/icons/Manbag.png musfirahdamla https://padlet.com/musfirahdamla/ANNweek7/wish/163017714 - Act as a supervisor, teacher
- Dataset is given to test]]> 2017-03-28 06:35:26 UTC https://padlet.com/musfirahdamla/ANNweek7/wish/163017714 musfirahdamla https://padlet.com/musfirahdamla/ANNweek7/wish/163018054 - outliers
- Have to find balance/tradeoff between the bias and the variance
- Issues of complexity - Need to have huge amount of data in order to get accuracy
- Dimensionality of input space - weightage, lots of hidden layers
- Noise]]> 2017-03-28 06:37:13 UTC https://padlet.com/musfirahdamla/ANNweek7/wish/163018054 musfirahdamla https://padlet.com/musfirahdamla/ANNweek7/wish/163019300 - No teacher
- Find hidden structure - gives dataset it will know it own ways through the layers
Used in:
- Data mining - Will find it own pattern. eg: knows when people will draw money the most
- Clustering, PCA (principle component analysis), ICA (Independent component analysis) , SVD (Singular value decomposition)

SOM (Self organising map) -- unsupervised learning technique(?) ]]> 2017-03-28 06:44:12 UTC https://padlet.com/musfirahdamla/ANNweek7/wish/163019300 musfirahdamla https://padlet.com/musfirahdamla/ANNweek7/wish/163020638 - Dynamic temporal behaviour - for inconsistent situation. eg: Recognizing  Handwriting - it's similar but slightly different.
- Need Feedforward , feedback

- Use gradient descent, hebbian learning]]> 2017-03-28 06:51:32 UTC https://padlet.com/musfirahdamla/ANNweek7/wish/163020638