2. RNN(Recurrent Neural Network)
1. RNN is a type of neural network
2. Mainly works on time series data. It is a collection of data in some timely order. Generally
the sequence is equispaced.
3. It has internal/hidden states to process the input sequence.
4. Output of a previous time step is also used in the next time step.
3. Practical Use of RNN
It is least explored but has high potential in technical ïŹeld.
Some of its uses are handwriting recognition, speech recognition, Stock Market
prediction etc.
11. LSTM(Long-Short Term Memory) Network
It is a type of RNN. it uses some additional units to the standard
units. A âmemory cellâ is used to store the information for a long
period of time.
LSTM consists of a cell and three gates viz, input gate(i), forget
gate(g), output gate(o).
12. LSTM contd..
Cell remember the information over time and Gates control the flow of
informations into and out of the cell.
f: Forget gate, Whether to erase cell
i: Input gate, whether to write to cell
o: Output gate, How much to reveal cell
14. Whatâs next
Furthering our knowledge of RNNs and timeseries prediction.
Getting a grasp of ïŹnancial stock market terminology and other relevant domain
expertise to understand the problems better.
Understanding how these are linked to bring out stock market predictions based on
machine learning techniques.