UW Introduction to Data Science Week1讨论加号召报名帖

小K 发表于 2013-5-13 11:11
eg. good =3, not good= -2, then "my day is not good" should count only -2?

啊啊~应该算-2吧，可是这么说我得代码把它算成3了啊。

这都通过了。。。

只有很少几个含空格的短语，我考虑先去match那几个短语，记下谁能被match
then split to words
如果split以后的单词不在短语key里， 就加上，
如果在短语中，则不添加

那就是brute force凡是含有的全部统统加上？

小K 发表于 2013-5-13 11:16
那就是brute force凡是含有的全部统统加上？

恩。我刚才去forum搜了下有关phrase的帖子发现大多数人都没考虑phrase。不过有个人的方法感觉不错：

I had an elaborate process of building a list of phrases from the file (phrases were identified by spaces), dealing with them in the tweet first, then getting rid of them to avoid double counting the words that make up the phrase. I suspect none of that is needed to pass this grader & suggest try just treating them like the other sentiment terms.

他先单独把phrase放到一个list然后在每个tweet里面找phrase，找到了再把该词从tweet里去掉，然后再计算剩下的分数。感觉这个approach更合理但grader貌似很松。。。这样不知道会不会反而错。

在这：https://class.coursera.org/datas ... t_id=4818#post-4833

明白了，make sense比我说的办法合理。。


对了regex不熟啊。。。。为什么下面这段找不到match呢？


p = re.compile(r'good')
m = p.match( 'not good' )
if m:
    print 'Match found: ', m.group()
else:
    print 'No match'

标点符号要怎么搞？don't like 就不会被识别成dont like呀。。。。

小K 发表于 2013-5-13 11:27
明白了，make sense比我说的办法合理。。

Python offers two different primitive operations based on regular expressions: match checks for a match only at the beginning of the string, while search checks for a match anywhere in the string (this is what Perl does by default).

所以用search应该可以吧~

小K 发表于 2013-5-13 11:28
标点符号要怎么搞？don't like 就不会被识别成dont like呀。。。。

提交作业时先不考虑phrase吧。。。grader不够成熟的感觉。。。

gnijuohz 发表于 2013-5-12 22:35
Python offers two different primitive operations based on regular expressions: match checks for a  ...

got it, thanks!!