/**
	 * learn training data, printing SSE at 10 of the epochs, evenly spaced
	 * if a validation set available, learning stops when SSE on validation set rises
	 * this check is done by summing SSE over 10 epochs
	 * @param numEpochs		number of epochs
	 * @param lRate			learning rate
	 * @param momentum		momentum
	 * @return				String with data about learning eg SSEs at relevant epoch
	 */
	public String doLearn (int numEpochs, double lRate, double momentum) {
		String s = "";
		String tmp = "";

		if (validationData==null) s = super.doLearn(numEpochs, lRate, momentum);
					// if no validation set, just use normal doLearn
		else {
			if(hasLearned)return "";								//if has learnt , do not proceed
			validationData.clearSSELog();							//clear the validations data SSE log
			for (int epoch = 0; epoch < numEpochs; epoch+=10) {		//for every epoch , step (10) (because we check the SSE every 10 epochs)
				presentDataSet(validationData);
				tmp=super.doLearn(10, lRate, momentum);				//perform 10 learning epochs in the train data
																	//present the validation data into the network
				if(validationData.getTotalSSE()>currSSE) {			//if this SSE rises from the previous one
					hasLearned=true;								//break and give info
					s+="Stopped after "+trainData.sizeSSELog();
					break;
				}
				if(trainData.sizeSSELog()%100==0) {
					s+=addEpochString(trainData.sizeSSELog())+":"+trainData.dataAnalysis()+"\n"; //add string with analytics
				}

				currSSE=validationData.getTotalSSE();				//remember this SSE
				validationData.clearSSELog();						//clear the validation log
			}
		}

		return s;											// return string showing learning
	}