I'm writing these two days2017 Annual work summary, What I learned in this year, A lot to say.
Towards the end of year, Finally, I finished learning the algorithm in Zou Bo's machine learning video, Otherwise, it's the four hundred dollars that I really need to spend. In order to understand his formula derivation, We have to put the high number, linear algebra, Probability and statistics again, Then when learning the algorithm formula derivation, we found that the derivative matrix distribution is the same as never seen before, We can only go back and nibble again. Countless times in the process, I doubted that my IQ was passed on to my son, I didn't leave anything for myself; I've doubted many times whether I studied statistics in University, Actually rightt distributionF There's no impression of distribution or anything; I doubted many times whether I had learned machine learning when I was a graduate student, At that time, Mao didn't think the least squares had anything to do with the maximum likelihood estimation; And then there's a lot of skepticism about my memory. I really need to go back home—— The new algorithm is over. I don't remember what happened to the last one, Even if it's something I thought I understood a week ago—— Typical bear breaking stick, Um. I can only admit that I am a bear. I'm glad I didn't have the courage to go home and be a housewife, I can only insist on being tough and cheeky, At last, I learned seven, eight, eight, As for the forgotten five five six six, You can only read your notes a few times before you forget them.
It is strongly recommended that students who are good at mathematics in university learn machine learning, You will find that mathematics is really amazing; It is also strongly recommended that students who are not good at mathematics in universities also learn machine learning, And tell me how you feel, To give me some comfort, It turns out that my IQ is normal, Let me have more courage to keep going, Thank you!
In order to learn machine learning, Learnpython, Although I used it beforejava Ofweka, But now it seems that all the people use itpython Do machine learning, You see, most machine learning classes usepython To make code, And it's said that it can be directly used in the production environment, Of course I have to keep up with the trend, What's more, what I learned is to do projects, not research; Look at you.python Those kits fornumpy,pandas,matplotlib,statsmodels as well assklearn And so on. From matrix calculation to statistics to visualization to machine learning, Have everything that one expects to find, So, you have topython! It just means, I have to learn. I have to learn, My poor memory!
Before that, Because I heardspark thanhadoop Learn when you have more cowsspark, In order to learnspark Learnscala, Online becausespark yesscala Write so usescala writespark Program comparisonjava,python More authentic, Of course I have to learn the most authentic. I just knew I had to learnpython, Why should I be so embarrassed, It's too late to cry. And now think about it, I don't remember at allscala What is it, My poor memory!
This year, usespark+kafka+redis Made a recommendation; Two simple models are made with random forest and autoregression; Then? No more.. It's like this year, I really don't think I've done anything, I often wonder if I will be sent home because of no performance—— God bless!! Buddha bless!! Doom Asylum!!!
New work plan for the new year.
Read a Microsoft articleAI Articles written by female engineers, call“ IntroductionAI, How to choose a down-to-earth position”. Inside the handleAI Position divided into Algorithm Engineer（ Data scientist）, Machine learning Engineer（ Commissioning Engineer）, Three positions of data manager. say concretely, Algorithmic engineers invent or optimize algorithms and can integrate theory with practice; Parameter adjustment engineers use mature algorithms to optimize parameters for engineering industrialization; Data manager, Familiar with the business, familiar with the data, able to simply analyze the data, find out the possible requirements, possible models, and coordinate the algorithm engineers or parameter engineers to complete the data products. There's a very funny analogy：
- The algorithm is to kill the dragon, Fight for battle, Beyond heaven;
- Engineering is hunting, Galloping horse, Drunkenness song;
- Making data is raising pigs, Mix pig food and clear pig manure every day, A face of dirt and a body of mud.
Um. Down to earth analysis of my current situation： Don't even think about Tu Long; The hunter will learn to shoot arrows on horseback, Maybe lucky enough to hunt a hare, a deer or something, Wolf, tiger, forget it, I'm afraid of being eaten; I don't want to raise pigs, But no one seems to be doing it for me. therefore, Down to earth,2018 My job in： Pig raising, Beat rabbits. Um. Also, Organize learning notes, summaries and so on into blog—— If you have time to finish housework with your child after work……
Okay, a new year, I hope I can survive the middle age crisis, Come on!