I got up around 115 pm to open the door for apartment service staff who came to repair the handle of the faucet in the kitchen which has fallen off for two times.

I guess I would stay in bed for another hour if he had not knocked the door so loudly. It wasn’t because I am lazy, though I know I am. It was because I didn’t sleep until 5 am in the morning for trying hard to understand some probability function.

Yes, I got some new knowledge of a python package called NGBoost by randomly browsing the website of Andrew Ng’s stanford ML group. I found this NGBoost package is very useful that it can output the probability distribution of certain regression prediction which makes the result more informative and colorful. But I have questions or I don’t understand thoroughly about the probability distribution so again I dive into the research of what does the distribution mean. One thing led to another, I end up with watching statquests. Luckily, I found some of the answers.
So that was what I did last night before I went to bed. Today, I fried 12 dumplings for breakfast. By the way, I definitely would buy frozen dumplings to fry back home. I really like this way of cooking dumpling rather than just boiling. After having breakfast, I finished reading 2 articles on reuters and the one on Gartner about the hype cycle of new technologies introduced by Chan’s NEJM article which I like very much. But when I want to reflect on last night’s understanding about probability distribution especially PDF, I met another problem. I don’t understand what y axis number stand for on a certain x because from statquest’s video Josh introduce the idea of likelihood and that is calculated from the PDF by x so “is y equals to likelihood?” that question puzzled me a lot making me scratching my head again. I have to dive into videos and web searching again. I even rewatched Josh’s video several times until I went to poop and check the question on Zhihu. Then it turns out that this question is very common and both chinese and western learners would raise the same confusion. Yet some guy made a very good analogy and explanation that helped me work the idea out. Actually the y on certain x means the density or, from my understanding, the amount of contribution to the cumulative probability from certain x and the likelihood is another concept about distribution choice or fitting although they come through the same calculation. I guess this is my biggest knowledge income today.

After having learned something, I decided to buy some gift for tommy in Marshall. What happened was I got myself a new shirt and short which only cost me $20. And I planned to give tommy the walkie-talkie as the gift. I am a bad daddy, right? I love you, tommy and louis and I miss you.

I did not do much after coming back. I like the shirt and short that led to me buying more on its official site where I found it was even cheaper online than in Marshall which made me feel bad, just joking. It’s only one dollar cheaper for a price tag of $7.99 in Marshall.

I don’t know maybe it’s because I had done too much thinking in statistics these days. I really don’t want to do more hard thinking into project design.
O.K. my plan is first revisit all the candidates ideas I drafted in pages and reorganize them into a concise and structured list such that I can have a clear road map of what to do next week, next month, next quarter, and next year.
I know it will be hard. Maybe I don’t know it would be this hard. OK. Can I finish the organizing of my research topic from the messy draft? Let’s try again tomorrow or hours later.