Rstudio- stock project
Budget: $30 – $250 SGD
This project aims to help you understand data analytics in a real-world context. You will apply various data analytic models to option pricing. The standard option pricing method (e.g., Black-Scholes) assumes the stock return follows a specific distribution and derives a formula for option prices. In this project, you will price options via a data-driven approach.
The dataset consists of observations of various option contracts on a stock traded in the U.S. market. Each observation consists of 16 variables:
1. Premium is the premium of the contract.
2. Type is the type of the contract-C for call and P for put.
3. Spot is the current price of the stock.
4. Strike is the strike price of the contract.
5. Maturity is the time till expiration, measured in years.
6. OI is the contract's open interest, i.e., the number of outstanding contracts.
7. Close1 – Close10 are the closing prices of the stock in the last 10 trading days;
Close10 is the closing price of the last trading day.
In practice, all these variables except Premium are easy to obtain. You wish to predict Premium based on other variables. The dataset has been split into a training set and a test set.
Tasks
The following tasks should be performed on the training set. You are not allowed to use the
test set.
Data pre-processing and descriptive study.
a. Provide a descriptive study of the dataset. Tables and figures could be helpful.
b. Handle possible missing values, outliers, etc.
Formulate the problem in the data analytic framework.
a. What is the response? What are the predictors? Are they quantitative or
qualitative?
b. How many observations and predictors are there?
c. Is this an inference or prediction problem?
d. Is this a regression or classification problem?
Apply data analytic techniques to solve the problem.
a. Apply various data analytic models (e.g., linear and nonlinear models, tree-based
models, etc.) to predict Premium.
b. Perform variable selection/model selection for each model (i.e., choosing
appropriate model complexity) via cross-validation or other appropriate methods.
IV. Fit each model with the whole training set and the selected complexity. Comment on your models.
Now you should have several models built from the training set. You wish to evaluate the performance of each model. The following task should be performed on the test set.
V. Evaluate the performance of different data analytic models.
a. Use your models to predict observations in the test set. Calculate the test MSE.
Comment on your findings.
b. What is the best model? What is the corresponding test MSE? Explain your
results.
VI. Write a report to demonstrate how you approached the problem and explain your results.
The report should be formatted properly, and it will be graded for both content and presentation. Please do NOT include any code in your report. Please limit the length of your report to 5 pages.
The dataset consists of observations of various option contracts on a stock traded in the U.S. market. Each observation consists of 16 variables:
1. Premium is the premium of the contract.
2. Type is the type of the contract-C for call and P for put.
3. Spot is the current price of the stock.
4. Strike is the strike price of the contract.
5. Maturity is the time till expiration, measured in years.
6. OI is the contract's open interest, i.e., the number of outstanding contracts.
7. Close1 – Close10 are the closing prices of the stock in the last 10 trading days;
Close10 is the closing price of the last trading day.
In practice, all these variables except Premium are easy to obtain. You wish to predict Premium based on other variables. The dataset has been split into a training set and a test set.
Tasks
The following tasks should be performed on the training set. You are not allowed to use the
test set.
Data pre-processing and descriptive study.
a. Provide a descriptive study of the dataset. Tables and figures could be helpful.
b. Handle possible missing values, outliers, etc.
Formulate the problem in the data analytic framework.
a. What is the response? What are the predictors? Are they quantitative or
qualitative?
b. How many observations and predictors are there?
c. Is this an inference or prediction problem?
d. Is this a regression or classification problem?
Apply data analytic techniques to solve the problem.
a. Apply various data analytic models (e.g., linear and nonlinear models, tree-based
models, etc.) to predict Premium.
b. Perform variable selection/model selection for each model (i.e., choosing
appropriate model complexity) via cross-validation or other appropriate methods.
IV. Fit each model with the whole training set and the selected complexity. Comment on your models.
Now you should have several models built from the training set. You wish to evaluate the performance of each model. The following task should be performed on the test set.
V. Evaluate the performance of different data analytic models.
a. Use your models to predict observations in the test set. Calculate the test MSE.
Comment on your findings.
b. What is the best model? What is the corresponding test MSE? Explain your
results.
VI. Write a report to demonstrate how you approached the problem and explain your results.
The report should be formatted properly, and it will be graded for both content and presentation. Please do NOT include any code in your report. Please limit the length of your report to 5 pages.