Question

我正在尝试实施候选消除算法＆＃34;在Python中，但我的代码不起作用。

我写了3个函数：

consistent用于检查假设和培训示例之间的一致性
more_general用于查找更一般的参数
more_specific用于查找更具体的参数

但我的算法不会添加或删除G和S中的假设。我找不到问题所在。你能救我吗？

# the general hypothesis

G = [ ('?', '?', '?', '?') ]

# the specific hypothesis

S = [('0', '0', '0', '0')]

# attributes:
AV = (['short', 'far'], ['cheap', 'expensive'], ['many', 'none'], ['yes', 'no'])

# training examples:
D = [

    {'sample': ('far',   'cheap',     'many', 'no' ), 'positive': True },
    {'sample': ('short', 'expensive', 'many', 'no' ), 'positive': True },
    {'sample': ('far',   'expensive', 'none', 'yes'), 'positive': False},
    {'sample': ('short', 'cheap',     'none', 'yes'), 'positive': False},
    {'sample': ('short', 'cheap',     'many', 'yes'), 'positive': True }
]




 def consistent(hypothesis, sample):

   return all([hypothesis[i] == sample[i] or hypothesis[i] == '?' for i in 
   range(len(hypothesis))])


 def more_general(a, b):

    result = False
    if a == '0' and b != '0':
       result = True
    elif a != '?' and b == '?':
       result = True

    return result


 def more_specific(a, b):

    result = False
    if a == '?' and b != '?':
       result = True
    elif a != '0' and b == '0':
       result = True

    return result


for d in D:

    if d['positive']:
        G = [g for g in G if consistent(g, d['sample'])]
        for s in S:
            if not consistent(s, d['sample']):
                S.remove(s)

                # Adding to S all minimal generalizations of s by h:
                dd = d['sample'] 
                if s == 0:
                    h = dd[s]
                else:
                    h = '?'  


                if consistent(h, d['sample']) and any([more_general(g, h) for g in G]):
                    S.append(h)

                #Removing from S any hypothesis that is more general than     another hypothesis in S
                for s2 in S:
                    if any([more_general(s2, s3) and not s2 == s3 for s3 in S]):
                        S.remove(s2)

    else:
        S = [s for s in S if not consistent(s, d['sample'])]
        for g in G:
            if consistent(g, d['sample']):
                G.remove(g)

               # Add to G all minimal specializations h of g
                for ai in range(len(AV)):
                    if g[ai] == '?':
                        h = list(g)
                        h[ai] = AV[ai][1 - AV[ai].index(d['sample'][ai])]
                        h = tuple(h)
                        if not consistent(h, d['sample']) and any([more_specific(s, h) for s in S]):
                            G.append(h)




    print('Sample: {} {}\nG: {}\nS: {}\n'.format('+' if d['positive'] else '-', d['sample'], G, S))

Answer 1

这是我的代码崩溃的地方：

        S.remove(s)

        dd = d['sample'] 
        if s == 0:
            h = dd[s]
        else:
            h = '?'  

        if consistent(h, d['sample']) and any([more_general(g, h) for g in G]):
            S.append(h)

s是一个元组，但在if条件下，与零比较时，您将其视为标量值。在if后，h包含标量，但在下一个if中，对consistent()的调用期望其第一个参数h为元组，不是标量。

我的候选消除算法不起作用

1 个答案: