An interface designed to abstract classes that store neural data. This interface is designed to provide NeuralDataPair objects. This can be used to train neural networks using both supervised and unsupervised training. Some implementations of this interface are memory based. That is they store the entire contents of the dataset in memory. Other implementations of this interface are not memory based. These implementations read in data as it is needed. This allows very large datasets to be used. Typically the add methods are not supported on non-memory based datasets.
For a list of all members of this type, see INeuralDataSet Members .
| Type | Description |
|---|---|
| BasicNeuralDataSet | Basic implementation of the NeuralDataSet class. This class simply stores the neural data in an ArrayList. This class is memory based, so large enough datasets could cause memory issues. Many other dataset types extend this class. |
| XMLNeuralDataSet | A data source that reads XML files. This class is not memory based, so very large XML files can be used, without problem. The XML data is assumed to look something like below. The names of the tags can be configured using the various properties on this object. <DataSet> <pair> <input><value>0</value><value>0</value></input> <ideal><value>0</value></ideal> </pair> <pair> <input><value>1</value><value>0</value></input> <ideal><value>1</value></ideal> </pair> <pair> <input><value>0</value><value>1</value></input> <ideal><value>1</value></ideal> </pair> <pair> <input><value>1</value><value>1</value></input> <ideal><value>0</value></ideal> </pair> </DataSet> |
| CSVNeuralDataSet | An implementation of the NeuralDataSet interface designed to provide a CSV file to the neural network. This implementation uses the BasicNeuralData to hold the data being read. This class has no ability to write CSV files. The columns of the CSV file will specify both the input and ideal columns. This class is not memory based, so very long files can be used, without running out of memory. |
| ImageNeuralDataSet | Store a collection of images for training with a neural network. This class collects and then downsamples images for use with a neural network. This is a memory based class, so large datasets can run out of memory. |
| MarketNeuralDataSet | A data set that is designed to hold market data. This class is based on the TemporalNeuralDataSet. This class is designed to load financial data from external sources. This class is designed to track financial data across days. However, it should be usable with other levels of granularity as well. |
| SQLNeuralDataSet | A dataset based on a SQL query. This is not a memory based dataset, so it can handle very large datasets without a memory issue. and can handle very large datasets. |
| TemporalNeuralDataSet | This class implements a temporal neural data set. A temporal neural dataset is designed to use a neural network to predict. A temporal dataset is a stream of data over a time range. This time range is broken up into "points". Each point can contain one or more values. These values are either the values that you would like to predict, or use to predict. It is possible for a value to be both predicted and used to predict. For example, if you were trying to predict a trend in a stock's price fluctuations you might very well use the security price for both. Each point that we have data for is stored in the TemporalPoint class. Each TemporalPoint will contain one more data values. These data values are described by the TemporalDataDescription class. For example, if you had five TemporalDataDescription objects added to this class, each Temporal point object would contain five values. Points are arranged by sequence number. No two points can have the same sequence numbers. Methods are provided to allow you to add points using the Date class. These dates are resolved to sequence number using the level of granularity specified for this class. No two points can occupy the same granularity increment. |
Namespace: Encog.Neural.NeuralData
Assembly: encog-core-cs (in encog-core-cs.dll)