Nowadays, people show more and more enthusiasm for applying machine learning methods to finance domain. Many scholars and investors are trying to discover the mystery behind the stock market by applying deep learning. This thesis compares four machine learning methods: long short-term memory (LSTM), gated recurrent units (GRU), support vector machine (SVM), and eXtreme gradient boosting (XGBoost) to test which one performs the best in predicting the stock trend. We evaluate Lithia Motors prediction models with Modular Neural Network (CNN Layer) and Spearman Correlation1,2,3,4 and conclude that the LAD stock is predictable in the short/long term. According to price forecasts for (n+1 year) period: The dominant strategy among neural network is to Hold LAD stock.

Keywords: LAD, Lithia Motors, stock forecast, machine learning based prediction, risk rating, buy-sell behaviour, stock analysis, target price analysis, options and futures.

## Key Points

1. Market Signals
2. What is Markov decision process in reinforcement learning?
3. Should I buy stocks now or wait amid such uncertainty? ## LAD Target Price Prediction Modeling Methodology

We present an Artificial Neural Network (ANN) approach to predict stock market indices, particularly with respect to the forecast of their trend movements up or down. Exploiting different Neural Networks architectures, we provide numerical analysis of concrete financial time series. In particular, after a brief r ́esum ́e of the existing literature on the subject, we consider the Multi-layer Perceptron (MLP), the Convolutional Neural Net- works (CNN), and the Long Short-Term Memory (LSTM) recurrent neural networks techniques. We consider Lithia Motors Stock Decision Process with Spearman Correlation where A is the set of discrete actions of LAD stock holders, F is the set of discrete states, P : S × F × S → R is the transition probability distribution, R : S × F → R is the reaction function, and γ ∈ [0, 1] is a move factor for expectation.1,2,3,4

F(Spearman Correlation)5,6,7= $\begin{array}{cccc}{p}_{a1}& {p}_{a2}& \dots & {p}_{1n}\\ & ⋮\\ {p}_{j1}& {p}_{j2}& \dots & {p}_{jn}\\ & ⋮\\ {p}_{k1}& {p}_{k2}& \dots & {p}_{kn}\\ & ⋮\\ {p}_{n1}& {p}_{n2}& \dots & {p}_{nn}\end{array}$ X R(Modular Neural Network (CNN Layer)) X S(n):→ (n+1 year) $∑ i = 1 n s i$

n:Time series to forecast

j:Nash equilibria

k:Dominated move

a:Best response for target price

For further technical information as per how our model work we invite you to visit the article below:

How do AC Investment Research machine learning (predictive) algorithms actually work?

Sample Set: Neural Network
Time series to forecast n: 22 Oct 2022 for (n+1 year)

According to price forecasts for (n+1 year) period: The dominant strategy among neural network is to Hold LAD stock.

X axis: *Likelihood% (The higher the percentage value, the more likely the event will occur.)

Y axis: *Potential Impact% (The higher the percentage value, the more likely the price will deviate.)

Z axis (Yellow to Green): *Technical Analysis%

## Conclusions

Lithia Motors assigned short-term Caa2 & long-term B3 forecasted stock rating. We evaluate the prediction models Modular Neural Network (CNN Layer) with Spearman Correlation1,2,3,4 and conclude that the LAD stock is predictable in the short/long term. According to price forecasts for (n+1 year) period: The dominant strategy among neural network is to Hold LAD stock.

### Financial State Forecast for LAD Stock Options & Futures

Rating Short-Term Long-Term Senior
Outlook*Caa2B3
Operational Risk 7031
Market Risk3843
Technical Analysis5444
Fundamental Analysis3166
Risk Unsystematic3035

### Prediction Confidence Score

Trust metric by Neural Network: 75 out of 100 with 725 signals.

## References

1. Candès E, Tao T. 2007. The Dantzig selector: statistical estimation when p is much larger than n. Ann. Stat. 35:2313–51
2. Chow, G. C. (1960), "Tests of equality between sets of coefficients in two linear regressions," Econometrica, 28, 591–605.
3. Dietterich TG. 2000. Ensemble methods in machine learning. In Multiple Classifier Systems: First International Workshop, Cagliari, Italy, June 21–23, pp. 1–15. Berlin: Springer
4. V. Borkar and R. Jain. Risk-constrained Markov decision processes. IEEE Transaction on Automatic Control, 2014
5. Mikolov T, Yih W, Zweig G. 2013c. Linguistic regularities in continuous space word representations. In Pro- ceedings of the 2013 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pp. 746–51. New York: Assoc. Comput. Linguist.
6. J. Ott. A Markov decision model for a surveillance application and risk-sensitive Markov decision processes. PhD thesis, Karlsruhe Institute of Technology, 2010.
7. Jacobs B, Donkers B, Fok D. 2014. Product Recommendations Based on Latent Purchase Motivations. Rotterdam, Neth.: ERIM