查看完整版本 : 求救:大學數elective 選讀Numerical analysis 定 stochastic processes ?

22point 2019-12-31 05:58 PM

求救:大學數elective 選讀Numerical analysis 定 stochastic processes ?

其實兩個都有興趣,但沒時間選晒黎讀,所以想問下其實邊個有用d , 搵工/升學

IWC-Rolex 2019-12-31 08:24 PM

一個elective,能讀得有多深入?邊有大所謂,喜歡邊科未邊科喇,能考得好重實際。

narius 2020-1-1 12:32 AM

It also depends on the actual materials. Just a title tells us nothing.


Plus, undergraduate classes really do not really do much aside from giving you the basic ideas, may be with a little practice.


So flip a coin. 

22point 2020-1-1 05:56 AM

Stochastic Processes I
Random vectors, multivariate densities, covariance matrix, multivariate normal distribution. Random walk, Poisson process. Other topics if time permits.



Stochastic Processes II
Markov chains in discrete and continuous time, random walk, recurrent events. If time permits, topics chosen from stationary normal processes, branching processes, queuing theory.



Introduction to Numerical Analysis: Linear Algebra
Analysis of numerical methods for linear algebraic systems and least squares problems. Orthogonalization methods. Ill conditioned problems. Eigenvalue and singular value computations. Knowledge of programming recommended


Introduction to Numerical Analysis: Approximation and Nonlinear Equations
Rounding and discretization errors. Calculation of roots of polynomials and nonlinear equations. Interpolation. Approximation of functions. Knowledge of programming recommended.






基本上呢到就係兩個課程內容

narius 2020-1-1 11:07 AM

Well .. since numerical method is basically linear algebra in that context .. take that first. You will need it no matter what you do (assuming you need more).


Approximation to nonlinear equations is very situational. Very simple method (like just a simple discretization with a grid) works if you have enough computational power to brute force it. 



A lot of the stuff done in stochastic modeling needs linear algebra anyway. To be fair, it depends a bit on the instructor. Take markov chain as an example. It is one approach to just show you all the math and proofs. It is quite another to show how you apply the concept in an application. Although I guess if it is taught out of a math department, it would be on the abstract side. 


What do you want to do career-wise? If you want a STEM career, this decision is kind of moot since you will need an advance degree anyway.

22point 2020-1-1 11:25 AM

I think I would like to be a financial analyst or data analyst

IWC-Rolex 2020-1-1 12:16 PM

[quote]原帖由 [i]22point[/i] 於 2020-1-1 11:25 AM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512479381&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]
I think I would like to be a financial analyst or data analyst [/quote]

單講一個選修其實無大意義,整體會否會加分,真係要睇埋其他,請人都唔會單單因為你有讀,無讀一科而有咁大影響,反而GPA相對更重一些。

22point 2020-1-1 12:20 PM

[quote]原帖由 [i]IWC-Rolex[/i] 於 2020-1-1 12:16 PM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512481286&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]


單講一個選修其實無大意義,整體會否會加分,真係要睇埋其他,請人都唔會單單因為你有讀,無讀一科而有咁大影響,反而GPA相對更重一些。 [/quote]
Thank you ,ching.畢竟GPA反映一個人醒唔醒

le123 2020-1-1 09:28 PM

Stochastic processes for sure, if u learn more financial maths, those calculation must involve time factor. Stochastic process provide u basic idea on handling time factor. If ur calculus and alegbra good enough, u can handle numerical analysis. 

22point 2020-1-2 06:44 AM

[quote]原帖由 [i]le123[/i] 於 2020-1-1 09:28 PM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512503896&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]
Stochastic processes for sure, if u learn more financial maths, those calculation must involve time factor. Stochastic process provide u basic idea on handling time factor. If ur calculus and alegbra  ... [/quote]
Thank you ching. 但係我直接揀Math of finance 呢個course 會唔會更好定stochastic processes? 呢到係content
The Mathematics of Finance course:
Introduction to the mathematics of financial models. Basic probabilistic models and associated mathematical machinery will be discussed, with emphasis on discrete time models. Concepts covered will include conditional expectation, martingales, optimal stopping, arbitrage pricing, hedging, European and American options.

竹劍 2020-1-2 07:22 AM

[quote]原帖由 [i]22point[/i] 於 2020-1-2 06:44 AM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512516849&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]

Thank you ching. 但係我直接揀Math of finance 呢個course 會唔會更好定stochastic processes? 呢到係content
The Mathematics of Finance course:
Introduction to the mathematics of financial models. Basic probabilistic mod ... [/quote]
如果打算靠financial mathematics搵食, 九成九九要MFE.
咁就不如先上stochastic process打好底, 點用落去finance度等master再讀都唔遲.

不過學樓上話齋, 可能睇下邊個course個prof派grade鬆手D可能仲實際. 畢竟numerical analysis同埋stochastic process既重要性相差唔太遠.

22point 2020-1-2 12:59 PM

[quote]原帖由 [i]竹劍[/i] 於 2020-1-2 07:22 AM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512517326&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]

如果打算靠financial mathematics搵食, 九成九九要MFE.
咁就不如先上stochastic process打好底, 點用落去finance度等master再讀都唔遲.

不過學樓上話齋, 可能睇下邊個course個prof派grade鬆手D可能仲實際. 畢竟numerical analysis同埋stochastic process既重要性相差唔太遠. ... [/quote]
都明ga 基本上人地都睇左gpa 先

narius 2020-1-2 02:21 PM

[quote]原帖由 [i]竹劍[/i] 於 2020-1-2 07:22 AM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512517326&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]
不過學樓上話齋, 可能睇下邊個course個prof派grade鬆手D可能仲實際. 畢竟numerical analysis同埋stochastic process既重要性相差唔太遠.[/quote]

Sigh .. i wish students care a little more about actual learning, than just a grade. But i do understand why people are so obsessed with KPIs.

If you go interview for a data science job (and at least in the US), they give you actual problem to solve. Having a high GPA does not necessarily contributes to that, particularly if you are in large classes where the instructor is too lazy to do anything but multiple-choice exams.

le123 2020-1-2 02:43 PM

[quote]原帖由 [i]22point[/i] 於 2020-1-2 06:44 AM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512516849&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]

Thank you ching. 但係我直接揀Math of finance 呢個course 會唔會更好定stochastic processes? 呢到係content
The Mathematics of Finance course:
Introduction to the mathematics of financial models. Basic probabilistic mod ... [/quote]
其實 financial maths 同 stochastic processes 係兩樣唔同既範疇,financial maths 計下 D bond, mean variance portfolio return, option pricing, CAPM,而  stochastic processes 係 part of probability which involves time factor, 如果你計一D adv. finance maths, e.g. ito lemma 咁你就需理解 stochastic processes

其實 finance 係好闊既 subject, 如果你係想做 Quant, financial eng, hedge fund, 咁其實stochastic processe係必須, 不過唔係讀完個 degree就了事, 起碼都 MSC, PHD 先可以入門。如果你係做 financial analyst, 咁其實係 target CFA, 或讀個MS finance. 而家 data analyst 係轉左做 data scientist, 做 big data, machine learning, data visualiazation, 咁係要識 statistics, data mining algorithms, 同 programming 野。

我會建議你再理解下自己想做咩範疇先。

22point 2020-1-2 02:53 PM

[quote]原帖由 [i]le123[/i] 於 2020-1-2 02:43 PM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512534430&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]

其實 financial maths 同 stochastic processes 係兩樣唔同既範疇,financial maths 計下 D bond, mean variance portfolio return, option pricing, CAPM,而  stochastic processes 係 part of probability which involves time f ... [/quote]
如果係咁睇黎Numerical analysis 都係基本要識,我都知係一定要讀上去,但想做幾年野先再讀。

le123 2020-1-2 03:38 PM

[quote]原帖由 [i]22point[/i] 於 2020-1-2 02:53 PM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512534839&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]

如果係咁睇黎Numerical analysis 都係基本要識,我都知係一定要讀上去,但想做幾年野先再讀。 [/quote]Numerical analysis, 如果你 basic maths 夠 solid, 其實睇下都識計。都只係 calculus & algebra.

IWC-Rolex 2020-1-2 04:51 PM

[quote]原帖由 [i]22point[/i] 於 2020-1-2 02:53 PM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512534839&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]

如果係咁睇黎Numerical analysis 都係基本要識,我都知係一定要讀上去,但想做幾年野先再讀。 [/quote]

如果真心想讀上去,絶對唔建議先做幾年,呢個唔同商科,如果屋企唔使樓主返工賺錢,應該即讀。直上PhD應該有有full scholarship連生活費,如果拿唔到scholarship,就要先考慮自己能力,合不合適繼續讀!
Master的話,值得讀既Master唔會太多。

竹劍 2020-1-2 08:53 PM

[quote]原帖由 [i]22point[/i] 於 2020-1-2 02:53 PM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512534839&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]

如果係咁睇黎Numerical analysis 都係基本要識,我都知係一定要讀上去,但想做幾年野先再讀。 [/quote]
真係想搞quantitative finance或者data science既建議趁早讀, no matter MSc or PhD. 唔讀隨時連行都入唔到.

[[i] 本帖最後由 竹劍 於 2020-1-2 11:23 PM 編輯 [/i]]

22point 2020-1-3 02:21 PM

我諗應該會master 先 ,但應該選讀CS? 定金融工程?

一個壞咗嘅城市 2020-1-3 02:42 PM

elective 數係冇用嘅。。。

如果你諗住畢業之後打份工。。。根本讀咩都一樣

竹劍 2020-1-3 05:14 PM

[quote]原帖由 [i]22point[/i] 於 2020-1-3 02:21 PM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512575902&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]
我諗應該會master 先 ,但應該選讀CS? 定金融工程? [/quote]
你問下自己讀完想做乜先?

而且CS其實最好bachelor讀, 果堆基礎野做野計重要過上advanced野. 有時間最好bachelor去上番三幾件CS course (e.g. OOP, data structure, computer organization)
至於之後係再深入咁讀computer system, 定係走去學多D AI, 或者data science野or whatever, 就又係問番自己讀完想做乜了.

22point 2020-1-4 07:49 PM

[quote]原帖由 [i]竹劍[/i] 於 2020-1-3 05:14 PM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512583398&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]

你問下自己讀完想做乜先?

而且CS其實最好bachelor讀, 果堆基礎野做野計重要過上advanced野. 有時間最好bachelor去上番三幾件CS course (e.g. OOP, data structure, computer organization)
至於之後係再深入咁讀computer system, 定係走去學多D AI, 或者data science野or wha ... [/quote]
其實一值都有d分唔清到底cs 同 data science 一樣? 點解想做data science 又係take cs ga course 其實係咪好大關係

竹劍 2020-1-4 09:19 PM

[quote]原帖由 [i]22point[/i] 於 2020-1-4 07:49 PM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512629091&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]
其實一值都有d分唔清到底cs 同 data science 一樣? 點解想做data science 又係take cs ga course 其實係咪好大關係 [/quote]
簡單咁講, data science其實就係搞數據, stat D 同類.
不過呢個年代D data太多 (e.g. big data), 或者D 方法計算起黎太複雜 (e.g. deep learning), 所以就要好好咁學CS, 去解決點樣計佢既問題.

narius 2020-1-4 10:08 PM

[quote]原帖由 [i]22point[/i] 於 2020-1-4 07:49 PM 發表 [url=https://www.discuss.com.hk/redirect.php?goto=findpost&pid=512629091&ptid=28790458][img]https://www.discuss.com.hk/images/common/back.gif[/img][/url]

其實一值都有d分唔清到底cs 同 data science 一樣? 點解想做data science 又係take cs ga course 其實係咪好大關係 [/quote]

Ok .. here is a very brief run down on the different "type" of data science.

Data science (or something called analytics, data analytics, or business analytics) programs can be run out of a CS department, or the IS side of a business school, or even an econ department. Because it is hot, everyone is jumping into the game.

Obviously each program varies ... but roughly speaking:

1) CS data science programs focuses more on the technology and "data plumbing".
2) B-school type programs focuses more on applications. The techniques can be exactly the same thing.
3) Sometimes econ department will market econometrics as data science (not completely unreasonable) .. and basically is stat modeling with econ reasoning.

So if you want to focus on the advancement of the tech, go for a CS program. If you are more apply, go for a business school program.

From my experiences (this includes both research and consulting on data science issues), just the techniques and understanding of them are not quite enough to be good in a business data science team. Typically you also need some domain knowledge. A simple example is clustering (in fact, i just had this discussion with a data scientist of a silicon valley company yesterday). Without business domain knowledge (i am talking about technical modeling of business processes, not just some "soft" knowledge of your business), it is very hard to formulate a reasonable distance measure for the clustering. The exercise always has to tie back to the final business objective. But *only* knowing the business is also not enough. For example, in this case one needs to be able to reason about math structures in high dimension.
頁: [1]
查看完整版本: 求救:大學數elective 選讀Numerical analysis 定 stochastic processes ?