hints

//for all epoch (maxepoch)
    	//Collections.shuffle(training,random) 
    	//for training set.size - all instances of training set
    	//list of attributes and list of classvalues
    	//on class values, do forward prop
    		//set all inputs of input node
    			//for attributes.size
    			//depending on index of attribute value, use it inside inputnodes and update that indexnode using the indexofattribute
    	//inputnodes.geti: currentinputnode
    	//setinput.attribute.get(i)
    	//update hidden nodes: 4 hidden loops, iterate through
    	//calculate output
    	//iterate through output nodes
    	//update output
    	
    	
    	//backward prop
    	//similar to way we set input for input node, set outputgradient for output node
    	//helper method setGradient in node
    	//above, we were updating input nodes, here do output nodes (setinput= setgradient), attributeval=classvalues.get(i)
    	//update values for outputnodes, hidden nodes, weights for output and hidden nodes (call functions from node class)
    	//find total loss(go through input set- make var loss, keep adding using loss method) and print