# Chin Wei's log

## 2023/02/14 meeting

- literature review on block
- how to divide stocks into block
- what is the nature of the stocks
- discuss on proposal defence
- well-prepared before 15/6
- discuss about mean-variance portfolio
- constrain optimisation

## 2023/02/22 meeting

- https://www.markdownguide.org/basic-syntax/
- present on the types of stocks.
- Three research papers were given to him for review

## 2023/03/08 meeting
- Present the paper titled" Sectoral portfolio optimization by judicious selection of financial ratios via PCA"
- Need to identify study all sectors or certain sectors
- Study the importance of different groups of financial ratio
-If study all sector, then use certain FR to filter out

- Divided sector and do filtering before optimization
- Try to prepare preliminary result
-Method :
Equal weightage
Steepest descent
Conjugate gradient
- Prepare FR for these result

## 2023/03/14 meeting
- Study 'Stochastic Gradient Descent Algorithm With Python and NumPy'
- Learn the coding from 'Portfolio Optimization in Python'
- Calculate gradient of objective function
- apply 2 stock to the python program, compare the results from the python with the hand calculation
- divided the stocks into block(by certain way)

## 2023/03/23 meeting
- Present Stochastic Gradient Descent Algorithm
- Discuss Gradient Descent details that matching to project
- Extract stock price details (python)
- View on python coding

## 2023/03/30 meeting
- Present literature review 'Risk, Return and Portfolio Optimization for various industries in the ASEAN region'
- discuss VaR / CVaR
- discuss project details
- Stage 1 - Equal Weightage for chosen industry
- Stage 2 - Optimization within industry
- the way to check data before using it

## 2023/04/20 meeting
- learn fundamental of steepest descent
- discuss about steepest descent method apply on my study
- discuss the python algorithm for S.Descent
- discuss work plan (from now until submit proposal defence)

## 2023/04/26 meeting
- Update on the experiment on equal weightage result and steepest descent result.
- The results are based on 3 stocks only. Been asked him to increase the sample stocks from 3 to 30 stocks.

## 2023/05/2 meeting
- Present 7 set of portfolio with 30 stocks each set
- Compare the expected return between equal weightage and steepest descent
- Discuss preliminary result part for proposal defence

## 2023/05/11 meeting
- Objective: to optimise expected return while diversify the investment
- Present literature review part
- Learn the skill for writing and present proposal defence

## 2023/05/24 meeting
- review on proposal defence documentation
- restructure the content
- adjustment on the algorithm

## 2023/05/29 meeting
- review on proposal defence documentation
- minor correction on the documentation
- discussion about proposal defense presentation

## 2023/06/16 meeting
- review on proposal defence presentation draft
- add more important content and delete redundant

## 2023/06/22 meeting
- review on proposal defence presentation slide
- minor modify on the slide
- finalise presentation slide
- Prepare for proposal defence

## 2023/06/30
Proposal Defence

## 2023/07/31 meeting
- discuss about my performance in proposal defence
- discuss and amendment on proposal defence documentation
- discuss about research direction

## 2023/08/10 meeting
- Study and present study outcome
- discuss on the study

## 2023/08/21 meeting
- Study and present study outcome
- discuss on the study
- discuss about research progress report

## 2023/08/28 meeting
- Study and present study outcome
- discuss on the study
- discuss signal vs noise and plan to include this in my project
- study Marčenko–Pastur distribution

## 2023/09/4 meeting
- discuss on abstract for colloquium
- study on step and algorithm of Marchenko-Pastur portfolio denoising
- construct Marchenko-Pastur denoising and add into preliminary result
- study on L1/L2 norm

## 2023/09/11 meeting
- modify on abstract for colloquium
- discuss algorithm of Marchenko-Pastur portfolio denoising(eigenvalue)
- review on the python code (Marchenko-Pastur)

## 2023/09/18 meeting
- review on the python code (Marchenko-Pastur)
- compare back the step with the textbook

## 2023/09/25 meeting
- review and modify on Marchenko Pastur python code
- study on Empirical view of Marchenko-Pastur Theorem

## 2023/10/26 meeting
- discuss on Empirical view of Marchenko-Pastur Theorem and its algorithm
- try to apply covariance and correlation in the same programme

## 2023/10/31 meeting
- discuss on Empirical view of Marchenko-Pastur Theorem
- discuss about full rank matrix

## 2023/11/6 meeting
- reconstruct covariance
- verifiying denoising data
- reconstruct MP and apply back to the previous work

## 2023/11/22 meeting

## 2023/11/29 meeting

## 2023/12/4 meeting

## 2024/1/12 meeting
- prepare descriptive statistic for data
- mean/ SD/ min/ max/ median(P50)
- use rtn.describe()
- kurtosis test on return - high kurtosis may yeild extreme return/loss
- probability of eigenvalue in graph must <1

## 2024/1/18 meeting
- plot skewness and kurtosis graph
- JB/ KS test - test for normality
- keep green line below the MP pdf bound
- different bin with different output(use figsize function to zoom in)

## 2024/1/26 meeting
- check difference between JB and KS test
- assumption
- output different or not
- pro and con

- call function to list normality of all stock
- similarity for stock filter by MP

## 2024/2/2 meeting
- Recalculate for new correlation. Should be some correlation of stock with itself approximately 1

- Use python and excel to compare

- Reshuffle stock location and compare the output