Scribd
Upload
2 views

Experiment_1_Data_Feature_Translation

The experiment aims to analyze a 16-year time series of returns from companies in the CSI 300 Index using Python or Matlab. Key tasks include documenting changes in constituents, analyzing the probability distribution of selected companies' returns, and assessing correlations between consecutive returns. A detailed report with runnable code and graphical outputs is required for submission by the specified deadline.

Uploaded by

Ashwaq Iarab02
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
2 views

Experiment_1_Data_Feature_Translation

The experiment aims to analyze a 16-year time series of returns from companies in the CSI 300 Index using Python or Matlab. Key tasks include documenting changes in constituents, analyzing the probability distribution of selected companies' returns, and assessing correlations between consecutive returns. A detailed report with runnable code and graphical outputs is required for submission by the specified deadline.

Uploaded by

Ashwaq Iarab02
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Experiment 1: Discovering and Describing Data Features

Business Data Analysis

1. Purpose of the Experiment

Analyze the 16-year time series data of the returns of companies included in the CSI 300 Index.

Learn and practice using Matlab or Python for describing data features and exploring potential

patterns in the data.

2. Experiment Content

The CSI 300 Index constituents undergo a "removal" or "addition" process based on a dynamic

adjustment every six months.

The "Task 1 Data Source" contains a list of CSI 300 constituent companies from January 2006 to

June 2022 (each half-yearly adjustment

recorded in a separate worksheet) along with their monthly returns.

Using Python or Matlab, complete the following data analysis tasks:

1. Extract lists of companies "removed" and "added" during each adjustment, and write these into a

CSI 300 Constituents Change Record Table, one page per six-month period.

2. Identify companies that remained part of the CSI 300 Index throughout the entire period. Select

five such companies, analyze the probability distribution of their per-share returns during this time,

and determine if it fits a normal distribution.

3. Further analyze the per-share returns of these companies. Assess whether the returns at time t+1

are correlated with the returns at time t.


3. Submission of Experiment Report

1. The experiment report should include runnable code (no screenshots), step-by-step explanations,

and the analysis results (use graphical output in TIFF format where possible; do not use

screenshots).

2. Submission Deadline: Follow the system's specified deadline in the CUPOST platform.

4. Additional Notes

1. First, consider, research, and experiment with how to programmatically read and consolidate data

from multiple pages into a unified time series.

2. Draw histograms to preliminarily determine if the data fits a normal distribution. Then test methods

in Matlab or Python for normality testing, such as the Kolmogorov-Smirnov test (kstest) or the

Lilliefors test (lillietest). Research and test these methods online if needed. See the illustrative

example below.

*Illustration 1: Comparison of Empirical and Model Normal Distributions*

(*Note: Data in the illustration differs from the experiment data.*)

3. For any company, plot returns at time t on the x-axis and returns at t+1 on the y-axis. Discuss and

explore any identifiable data features.

*Illustration 2: Correlation of Stock Returns at t and t+1 for a Listed Company*

You might also like