Fix typo of sum_squared_error at common.functions

[NicholasWon47]

wrong expression for sum_squared_error(mean_squared_error) function

[koki0702]

Thanks for PR! You're probably right.

• Mean squared error is np.sum((y-t)**2)/len(y) .
• Sum of squared error is 0.5 * np.sum((y-t)**2) (or just np.sum((y-t)**2)).

I made a mistake, so I fixed.

https://github.com/oreilly-japan/deep-learning-from-scratch/wiki/errata#%E7%AC%AC14%E5%88%B7%E3%81%BE%E3%81%A7

Sorry to confuse you.