为什么基于神经网络的分类器比贝叶斯网络更好？
我正在尝试找到一种很好的分类方法来解决我的问题，即将多个丢失、截断或错误数据值的客户记录分类为不同的客户类别，即对一个或多个客户记录进行分类，看看它是否属于同一客户或不同的客户顾客。为什么我应该使用神经网络而不是贝叶斯网络？我的教授说神经网络是最好的方法。
I am trying to find a good classification approach for my problem of classifying multiple customer records with missing, truncated or wrong data values into different customer categories i.e. to classify one or more customer record and see if it belongs to the same customer or to a different customer. Why should I use a neural network for this and not a bayesian net? My professor said that a neural network is the best approach to it.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
这在很大程度上取决于您尝试分类的数据类型。神经网络通常擅长连续数据，而贝叶斯网络往往更适合离散数据。当然，连续数据可以通过将其放入存储桶中来离散化，但这是您可能不需要的另一层复杂性。
这两种方法（理论上）都能很好地处理丢失、截断和不正确的数据。
我建议你问问你的教授为什么他们认为神经网络是更好的方法。
It depends very much on the type of data you are trying to classify. Neural networks are typically good at continuous data whereas bayesian nets tend to work better with discrete data. Of course, continuous data can be discretised by putting it into buckets, but that's another layer of complexity that you may not need.
Both approaches (theoretically) cope well with missing, truncated and incorrect data.
I'd suggest that you ask your professor why they think a neural network would be a better approach.