Now on the numeric analysis of the class, all using Python, the actual working time?
Reply content:
For different quant posts, the software used varies greatly.
In the case of pricing models, it is possible to use MATLAB most often, because the language is simple and the expansion package is excellent. It is best suited to pricing derivatives that do not require time requirements.
If you are doing high-frequency trading, the most common is C + +, because it is fast. There are a lot of high-frequency companies, have their own write the bottom package, and for this spent a lot of manpower and material to maintain.
If you are writing a profit model, you will choose R more. There are many leading statistical model packages in R. Very powerful for statistical analysis and time series processing. But the expansion pack does not have the stability of MATLAB.
For Python, it's more like an almighty programming need. Whether it is network capture, statistical analysis, or numerical analysis, there is a good plan. However, in the area of specialization, it is weaker than the language mentioned above.
In fact, in different financial companies, have different language use of traditional, some like C + +, some like C #, and some like Java. But for a mature company, there is a need for good support for these languages. After all, a good trader is very rare, is it because of the different use of language to refuse to make him rich talent?
In fact, software is only a carrier, thought is the most core. I work in a place where everyone uses VBA. When trading, the full screen of Excel. As long as you can finish the work, who cares what language you use. First come to the conclusion: tools and language are not the focus. Mathematical ability and market understanding is.
I have been to several international major lines, the mainstream tools have
1. C + + on the company's own platform
2. The company's own platform of Python
3. Matlab
4. R
Among them, 1 and 2 emphasize autonomous platform mainly because its important library and part of the grammar are custom, basically can be regarded as half independent language. It doesn't make much sense to study outside. The company's high-level yard, a large heap of quant, a major ability is actually a mathematical ability, other can write on it.
The same is true for large funds. Small fund What is the line, VBA has. Answer the questions of the main question first,
in the field of financial engineering, Python is not only used, the most, but also more important .。 Reason: As a dynamic language Python, the language structure is clear and simple, the library is rich, mature and stable, scientific calculation and statistical analysis are very good, production efficiency is much higher than c,c++,java, especially good at strategy backtesting.
There are some exceptions:
1, in the past formed other language-based systems, such as the platform to write C + +, the API is C + +, the elderly are skilled in C + +, then they are mainly C + +.
2, some institutions to study mainly, perhaps they use matlab,r.
3, C + +, JAVA, VB and other languages of genius coder, I can only worship at a distance, you use the assembly is also higher than my efficiency 10 times times, you are my mind of God.
Describe yourself again, according to the classification, I should calculate a trader, a will write the program of the trader. My job is to analyze data, discover patterns, develop and validate strategies, then trade, and then optimize adjustments, reciprocating loops. We use automatic trading, but high-frequency trading is not my range.
I am personally moving from C-+ + to Java and then to Python for no reason, efficiency. People have limited energy, write procedures more time, trading money less time, my team also very much agree with my idea.
The reason I prefer Python:
1, right is efficiency. The more complex, lower-level language, the higher the error rate, the bug leads to data analysis errors, the trading strategy program crashes, light missed trading opportunities, heavy trading strategy itself leads to loss, the feeling of crying without tears ...
2, Python heavy development efficiency, imagine the same trading strategy, I have completed backtesting, to report to the boss to apply for the transaction amount, and you are still debugging C + +, the Boss will not think you are diligent, the boss would think you are some kind of animal, and then go to think about, such as the end of our red envelope thickness ....
3, Python performance is not bad, Python or call C library, and then consider PVM, in fact, and C performance differences can be negligible. Really care about the small difference in performance, it is recommended to spend 20,000 bucks to buy the latest Mac PRO, direct seconds kill. What did you say? The unit does not give the budget? For pleasure, can you do it at your own expense?
4, the language is not the main performance bottleneck, network, equipment, such as the impact of greater, such as C + + program faster 2us, and network delay consumption of 1ms.
I am familiar with the peers, understand Python, have the ability to go to Python, but also part of the programmer's decision, not to go, you understand.
A few exceptions to the front, this part as entertainment, does not belong to the content:
The first kind, is the kind that has the accumulation of the history, say the good-sounding is the knowledge base, said the bad hearing that is the historical burden. Most of the strategies on the old platform have been invalidated, and should have been pushed back again.
The second, those who sell the analysis report, those who use Matlab,r, may be an intern, and even never traded, his report is reliable. Further, if the report is reliable, why does he not trade himself, recreate a renaissance, why should I tell you that he is a big fool? Or is it stupid to buy a report?
Two days did not come, the attention of more people, clarification. Hair This post late at night just add finish class, in the car phone hastily dozen, words hasty, to brother disrespect, first apology.
A-share market hot, these days are working overtime, just woke up, said the wrong please forgive. There are several good implementation strategies this year, if you are interested, I can share two, of course, or discuss the application of Python in financial engineering mainly.
======================================================================
Originally written in this:
First said ranked first, a look at you is a layman, others ask Python, if you understand, directly with people say Python, do not know also to install force, know that the net you this kind of goods.
There are those who say that the tool is not important, you simply do not use the computer, go back with paper and pen line, as long as the rest of the life can run a strategy, you can run two a lifetime, do not worry, slowly.
As a dynamic language Python, the language structure is clear and simple, the library is rich, mature and stable, scientific computing and statistical analysis are very good, production efficiency is much higher than c,c++,java, especially good at Strategy backtesting, is the first choice of the big line.
What era, a PC performance is far beyond the past of the mainframe, who have kung fu to you c and C + +, know blind BB. Look at the article 5 years ago.
Python promises to be a financial language
http://www. infoq.com/cn/news/2010/ 04/python-abs
The financial industry uses PYTHON,R to do the research is more common. Python can do a lot of things with the scientific data calculation module. If you look at the entire U.S. side of the quant industry, quant use Python to apply a very high proportion, one of the three or four mainstream languages. Referring to the words of a PhD sister in mathematics, Python is the second generation language relative to C + +. Engineering children may use MATLAB More, while medical statistics are used in SAS. In fact, seriously learn, do projects, learn quickly one months familiar, three months proficiency.
Programming language is a computer-based language that is invented in English + symbols + mathematics. If these three you are well-mastered, just start learning it, do not tangle ... The core ability of Quant is what to learn quickly and grasp quickly. Come on! Above ~ First said ranked first, a look you are a layman, perhaps you have involved in one or two industries, but in terms of finance, you do not pretend, know that you are the goods.
Different business types, different functional parts, different company cultures, different team skills, determine what language the business is developed, what hardware architecture to use, what database to choose, and what development mode to choose.
Python is quick to get started, development speed, high productivity, still have its force can not and place, do not understand the blind BB. Software and tools are not the focus, more training will be able to get started, the key is the mathematical basis and application capabilities
