encog-java-core
encog-java-core copied to clipboard
jeffheaton
On high-res monitors, the Encog workbench and some GUI examples appear very small on high res computers. Not entirely sure what the solution is, but I am adding an issue...
BayesianEvent `public boolean hasChildren()` returns the value of `hasParents()`. I believe it should be `return this.children.size();` https://github.com/encog/encog-java-core/blob/5b9d9a8c8b4299b49ef574f2743d21d5a9438a6c/src/main/java/org/encog/ml/bayesian/BayesianEvent.java#L166
Got some unexpected results from ReLU activation function with custom `threshold` and `low` parameter values set. This simple patch should fix it, or I misunderstood something. Looks like same issue...
The FactorHyperNEATGenome and HyperNEATCodec classes do not implement Serializable. Simply adding Serializable to the class specs makes serialization work.
Dear Mr. Heaton, I am developing an application on the time series economic analysis and would like to use TemporalMLDataSet with 10 attributes to store my data. However, I found...
1. In Workbench, create .ega file for NN regression - using data: http://dione.zcu.cz/~toman40/encog/data9.zip 2. Execute task-full and "Stop all commands" when task-train begins. 3. "Normalize to training" `data9_eval.csv`. 4. Double...
There is a number of places where manual array copy is done instead of using System.arraycopy(), which is much faster (the bigger the array, the faster: ~2x for m=10, ~15x...
Now that Issue #28 has been added, it would be very nice to be able to add regularization, and other strategies to the analyst training. This should be added via...
Hello, At the moment, it seems that the implementation of the NEAT network only allows a fixed size input to be computed, and the network is reset afterwards. NEAT networks...
I am new to encog and try to fit the examples to my needs. I am using the SunSpotTimeseries in Java. By default, the lagWindowSize is 3 and the leadWindowSize...
