Abstract:

The housing market bubble burst caused the financial crisis of 2008. This caused the people of the US to suffer. A recession can have a big impact on a nation’s economy or the global economy. The 2008 housing recession had the worst impact not only on the American economy but also the global economy after the Great Depression. Many well-known statisticians and media houses were talking about a recession in 2020, not only that the housing prices were declining at the starting of 2019. This made me look into the housing market to see whether the price dip can cause a recession or is it that the housing market was just correcting itself from the 2018 price hikes. To do the analysis and prediction, I would be using the historical housing price data for California from 1990 to 2020. I would do a comparative study of three Machine Learning techniques these are Linear Regression, Hidden Markov Model and Long Short-Term Memory and report the result for each of them.

CS297 Results

• Data cleaning and preparation
• Coded HMM algorithm from Prof. Stamp’s book to know the working of the HMM algorithm
• Created an LSTM model for a smaller data set to learn how train an LSTM network
• Used Linear Regression on the initial housing data set

Proposed Schedule

Key Deliverables:

• Collecting data and implementing Linear Regression on the collected California housing dataset. The link for the dataset is mentioned in the reference. The data collected will have a median housing price for each month for all California counties.
• Build and train HMM to predict the housing market using the dataset stated above using Sci-Kit learn. The data will have states to help predict the housing price.
• Build and train LSTM using TensorFlow to predict the housing market. The LSTM network will build on the same dataset described above. It will predict housing prices from the given prices.
• CS 298 report.
• CS 298 presentation.

Innovations and Challenges:

• Collecting housing dataset that contained enough data for ML training and design was a challenge.
• Building and training LSTM network was time consuming
• HMM had not been used in housing market prediction. It had previously been used in stock market prediction

References:

[2010] Y. Demyanyk and I. Hasan, “Financial crises and bank failures: A review of prediction methods”, Omega, vol. 38, issue 5, pp.315-324, 2010.

[2017] E.J. Schoen, "The 2007–2009 Financial Crisis: An Erosion of Ethics: A Case Study", J. Bus. Ethics, vol. 147, pp. 805-830, Dec 2017.

[2006] M.R. Hasan and B. Nath, “Stock market forecasting using Hidden Markov Model: A New Approach”, 5th Intl. Conf. on Intel. Sys. Design and Appl., IEEE, 2006.

[2019] Y. Hu, X.Sun, X. Nie, Y. Li and L. Liu, “An Enhanced LSTM for Trend Following of Time Series”, IEEEAccess, IEEE, 2019.

[2008] M.G. Crouhy, R.A. Jarrow and S.M. Turnbull, “The Subprime Credit Crisis of 2007”, J. of Deriv, pp. 81-110, 2008.

[2016] R.Nyman and P.Ormerod, "Predicting economic recessions using machine learning algorithms", Dec 2016.