A financial company hires your team in order to develop the back-propagation neural network(s) in order to predict the next-week trend of the five stocks (i.e. go up, go down, or remain the same). In meantime, the company also offers you with the data for each stock in past 15 years. Every data record contains 20 attributes (like index values, earnings per share, revenues,capital investment, and so on). Provide answer for the following questions:

1) The team member A proposes that you must develop a single neural network which may handle all these stocks. However the member B insists that you have to design the five networks (one for each stock). Whom do you think is correct and why?

2) During the training, at a particular point, you notice that error rate (of each training cycle) has been oscillating (i.e. it reduces in the round n, increases in the round n+1, and decreases again in round n+2, and so on). Specify the reason for this phenomenon and what you must do about it?