Some Thoughts on sharing
Before learning through written several machine-related blog, to see the partial simplest form more easily accepted by everyone, but some of the more difficult projects, share effect is very general, because of the following aspects:
- The more difficult the project requires the project background, data forms, contest information, a certain understanding, otherwise easily confused, even to doing can not read;
- Limited individual's writing ability, including the overall structure of the blog is also not reasonable, not been able to speak to understand concise;
- Machine learning for understanding is not enough, so after a written blog fought a discount, a return to people do not understand is normal;
Currently considering how to organize my personal machine learning in the learning process, competitions and other projects into a series, that can help them deepen their understanding, if you could play for the slightest little friends to help, that's fine;
But this series is not considered good form to share, it is necessary to be able to give you some gains, but also show some advantages in my personal situation, I do not really understand the part of said pass on to you, not to mislead you, this is a bit difficult ah. . . .
But one thing I feel good, before seen in an article, saying that some foreign kaggle bigwigs to discuss with others when there is a good habit to share, not only that he would share his last resort, effective data handling, and he will try to share some of his previous passable, like in the good habit he shared his process of thinking a whole, not just a result of doing this feature works when it is very important that we see a good machine learning models, it is also very easy to copy reproduce its effect, but just for a problem they feel that they can not do, even if similar problems, and this is because the model is the result but how to reach this result, the process is more important, which contains a lot of thinking, try, do not know the process has become not why, I think this can be learned;
If someone can see this essay, I hope to give me a message, usually written for others to see things not many, so also some confusion in this area, but also brings hope to share some of their own things to discuss with you, along with progress , thanked him;
Written in 2020-01-21 23:47, Shenzhen;