What tools should I use for analyzing the trend or pattern of timestamped data?

1 次查看（过去 30 天）

Ran Wei 2018-1-8

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/376193-what-tools-should-i-use-for-analyzing-the-trend-or-pattern-of-timestamped-data

评论： Matt Sprague 2018-1-11

Hi folks,

I am currently doing a project in which I have timestamped data and maintenance report record. I want to find the connection between the data and the record. I tried various visualization methods but the result is not good, so I moved on to try machine learning.

My plan is to select time segments and label them with maintenance records. But the difficult is that one input is numerical and the other is not.

What is the best tool for this? Please comment if you have any suggestion.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Matt Sprague 2018-1-11

You may find this link helpful: Supervised Learning Algorithms: Categorical Predictor Support

Since I understand you are working on a combination of numerical and non-numerical data you might have some success with decision trees, SVM or a naive bayes. As far as which method is "best", that depends on the content and relationships present within your data and what you are trying to predict. I'd encourage you to try out each of these methods and see which fit your problem best! The Classification Learner and Regression Learner apps could help out with this since they would allow you to test and visualize the results across different methods quickly. Good Luck!

请先登录，再进行评论。

请先登录，再回答此问题。