[mlpack] Some ideas about the project String Processing Utilities

[Tan Zhiren]

Hello!


I am Zhiren, a University of the third grade students, majoring in computer science and technology, I found a surprising item in the browse the mlpack project, because it was I remain perplexed despite much thought of the problem, yes, it is String Processing Utilities.


Not long ago, I was doing a project with machine learning, a mobile node path is the input data, and other data is not the same, it is a string of characters, the composition of the place obviously, place order is very important information, so how to make the computer can accept this the data and understand it? The direct input of a string of characters is obviously an evasion of the problem. I checked some information did not find a solution, then I thought of a solution: the training data set I is a matrix, then I put the path string into a single location, such as schools, hospitals, supermarkets, which turns out to be a data item, I put it into the three so, since each data item is a clear meaning, which at the time can be said to be a helpless way, because of direct calculation is very complex character. So you can't imagine how excited I was when I saw this project, because I saw mlpack formally put forward this problem, and I had the chance to solve it, and I freed the later people from this trouble.


I am very confident to complete this project, I can skillfully use C++, and use it to complete several relatively large projects, the other I think the problem description to convert characters into digital computing is indeed a good idea, but I think that there is room for improvement, for example, we can get rid of the character of no great importance (is, are), and how to use simple numbers (0, 1) to express more information? Can this conversion be returned accurately?


At present, I have been familiar with some basic tutorials of mlpack, and have compiled some simple projects. If I want to apply for this project, what should I do next?

Look forward to your reply.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://knife.lugatgt.org/pipermail/mlpack/attachments/20180302/9babfeef/attachment.html>