python - 我如何检查我的聊天机器人中的 2 个句子是否相似

Question

我做了最简单的聊天机器人。它可以根据您之前希望它回答相同问题的内容来回答您的问题。代码有点像这样：

question = []
answer = []
qcount = 0
stop = 0

b = 0

while stop == 0:
    b = 0
    q = input("Ask me anything: ")
    if q == "stop":
        exit()
    for i in range(qcount):
        if q == question[i]:
            b = 1
            print(answer[i])


    if b == 0:
        question.append(q)
        qcount = qcount + 1

        a = input("How should i answer that? ")
        answer.append(a)

有没有转的方法

if q == question[i]

至

if q is similar to question[i]

?

标签： pythonartificial-intelligence

Answer 1

Aryan Mishra 已经为您提供了答案。我有类似的答案给你。你也可以试试这个。您不需要计数器和退出语句。您可以将 while 语句本身定义为守门员。

我又做了一些改进。虽然这不会给你一个完美的聊天机器人，但它更接近了。

question = []
answer = []

q = input("Ask me anything: ")
while q.lower() != 'stop':
    i = -1
    z = q.lower().split()
    z.sort()

    for x in question:
        y = x.split()
        y.sort()
        if all(elem in y for elem in z):
            i = question.index(x)
    if i >= 0:
        print(answer[i])
    else:
        question.append(q.lower())
        a = input("How should i answer that? ")
        answer.append(a)
    q = input("Ask me anything: ")

输出：

Ask me anything: What is your Name
How should i answer that? Joe
Ask me anything: What your name
Joe
Ask me anything: name
Joe
Ask me anything: your name
Joe
Ask me anything: what name
Joe
Ask me anything: what is name
Joe

正如你所看到的，当你问“什么是名字”时，它仍然假设你在问你的名字是什么。您需要使用它来获得更复杂的机器人。希望这可以帮助您朝着正确的方向前进。

我之前的答案也发布在这里。由于我们将字符串与列表进行比较，因此它必须完全匹配。检查有问题的 q 并没有真正给你带来优势。您将需要拆分单词并进行比较。这就是我在新回复中所做的（见上文）

question = []
answer = []

q = input("Ask me anything: ")
while q.lower() != 'stop':
    if q.lower() in question:
        i = question.index(q.lower())
        print (answer[i])
    else:
        question.append(q.lower())
        a = input("How should i answer that? ")
        answer.append(a)
    q = input("Ask me anything: ")