Walt Disney Company Stock Forecasting Using Linear Regression

pdf

School

University Canada West *

*We aren’t endorsed by this school

Course

502

Subject

Economics

Date

Apr 26, 2024

Type

pdf

Pages

Uploaded by JusticeIceBaboon53

1 Data Collection and Descriptive Statistics of the Walt Disney Company Phase - 3 University Canada West MBAF 502: Quantitative Reasoning and Analysis S-37 March 25, 2024 5.1 What method was used for forecasting? Linear regression analysis was used to forecast the stock price of Walt Disney and Ne:lix companies. The technique of regression analysis allows us to examine the rela@onship or correla@on between the dependent variable (stock costs), and the independent variable (@me period with days and months). The linear trend obtained by regression analysis allows us to assess the nature of the rela@onship between two variables and predict future values. The main indicators in forecasting using regression analysis are the strength, direction and reliability (validity) of the relationship. The strength of the relationship is determined by the absolute value of correlation varies from 0 to 1 (Pearson correlation coefficient). The direction of

2 the relationship is determined by the sign of the correlation coefficient: a positive coefficient - the relationship is direct; a negative coefficient - the relationship is inversed. Reliability of the relationship is determined by the p-level of statistical significance (the smaller the p-level, the higher the statistical significance or reliability of the relationship). 5.2 What were the forecast results? The results of the forecast were the predicted values of the companies' shares for three months - December 2023, January 2024 and February 2024 in Canadian dollars. 5.3 How did the forecast results agree with the actual historical data? 5.6 Were the correlations statistically significant? The forecast values differ from the actual values in both companies, and for the Walt Disney Company, the difference is very significant. The forecast values of Walt Disney's stock price. Table 1 Summary Output Walt Disney

3 The forecast of Walt Disney Company's stock value should be considered unreliable because the correlation between the forecast and actual values of prices is low at 0.48 (figure 1). Figure 1 y = -0.1517x + 94.083 R² = 0.481 $80 $81 $81 $82 $82 $83 $85 $90 $95 $100 $105 $110 $115 Predicted price Actual price Scatterplot Actual/Predicted Walt Disney

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

4 The correlation or R square between the share price and time (day and month) is weak and is only 0.55 (table 1). According to the figure 2, a linear dynamic in the direction of decrease in the value of shares over time. The stock value has been gradually decreasing since the beginning of 2023, but from the 7 th of November the stock values increased. This upward spike in value was not included by linear regression in the accounting, continuing the downward trend line to lower value (figure 2). Figure 2 Figure 3

5 Despite the discrepancies in the forecasted and actual values of stocks, the significance of the linear regression model is confirmed by the values of p- and F-coefficients. F (1.3584E-35) is less than a (0.05), indicating the presence of regression (table 1). This was indeed the case from the beginning of 2023 the value gradually decreased. The other parameter P-value which if less than 0.05 means that the coefficients have an effect on stock price and this regression is good enough. According to ANOVA analysis in the regression, our P-values of coefficients (Intersept 3.50165E-37 and slope 1.3584E-35) are much smaller than 0.05. Thus, the model does meet the linear dependence. The forecast values of Netflix's stock price. Table 2 Summary Output Netflix $60 $70 $80 $90 $100 $110 $120 15-11-22 03-02-23 24-04-23 13-07-23 01-10-23 20-12-23 09-03-24 The Walt Disney Company Fact Close Stock Price Walt Disney (Feb/2023- Nov/2023) Forecast Average Close Stock Price Walt Disney Forecast High Close Stock Price Walt Disney Forecast Low Close Stock Price Walt Disney Fact Close Stock Price Walt Disney (Dec/2023- Feb/2024) Linear (Fact Close Stock Price Walt Disney (Feb/2023-Nov/2023))

6 Figure 4

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

7 According to the results of regression analysis of Netflix’s share prices, the correlation (Pearson coefficient) between forecasted values and actual values is good and amounts to 0.75 (figure 4). It shows similarity of forecasted values with actual values. However, the value of approximation for the dependence of stock price and time variables is only 0.55 that indicates a weak relationship (figure 5). Nevertheless, the F-test and p-values are less than a (0.05) and confirm that the trend of stock price over time corresponds to a linear regression positive model (figure 6). Figure 5 y = 0.2119x + 365.41 R² = 0.7547 $450 $455 $460 $465 $470 $475 $480 $485 $490 $495 $400 $450 $500 $550 $600 $650 Predicted price Actual price Scatterplot Actual/Predicted Netflix

8 Figure 6 y = 0.4591x - 20323 $100 $150 $200 $250 $300 $350 $400 $450 $500 24-01-2023 15-03-2023 04-05-2023 23-06-2023 12-08-2023 01-10-2023 20-11-2023 09-01-2024 Fact Close Stock Price Netflix Fact Close Stock Price Netflix Linear (Fact Close Stock Price Netflix) $60 $160 $260 $360 $460 $560 $660 15-11-22 03-02-23 24-04-23 13-07-23 01-10-23 20-12-23 09-03-24 Netflix Fact Close Stock Price Netflix (Dec/2023- Feb/2024) Fact Close Stock Price Netflix (Feb/2023- Nov/2023) Forecast Average Close Stock Price Netflix Forecast High Close Stock Price Netflix Forecast Low Close Stock Price Netflix Linear (Fact Close Stock Price Netflix (Feb/2023-Nov/2023))

9 5.4 How were the correlations calculated? The tool Excel Data Analysis was used for the forecast. The data of two variables were used for calculation: independent variable (time -days and months) on the X axis and dependent variable - the stock price, which on the graph is located on the Y axis. The resulting tables contained residuals, R square, standard error, number of observations, coefficients (intercept and slope of equation), and as well as graphs. 5.5 What were the correlations between the daily returns? Based on the information provided, it appears that the correlations were not statistically significant. Despite having a Significance F value of less than 0.05, indicating overall significance of the regression model, individual components such as the p-value for the intercept (0.6477) and the value of the approximation close to zero suggest a weak relationship and lack of correlation between returns. Therefore, the correlations were not statistically significant. 5.7 Write the conclusion about the reliability of forecasting and correlation analysis as applied in this project. Thus, both companies show a linear trend in stock price changes over time, which is confirmed by Pearson correlation (more than 0.5), F-test and p-values (less than 0.05). To calculate the regression linear equations, the sample from February 2023 to November 2023 was used, and the forecast period was December 2023-February 2024. In this case, in the Walt Disney Company samples (or data) the stock prices decreased (negative correlation) and in the Netflix samples (data) it increased (positive correlation). The

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

10 trend should have allowed determining the main direction of the stock price movement, which is upward or downward. Nevertheless, the resulting predictive trend for the Walt Disney Company showed that the values deviate significantly from the real values and have an inverse direction. In the forecast period, the trend has changed - the stock price began to rise, not fall (figure 3). For the Netflix Company, the forecast values retained a positive trend - growth in the value of shares over time, but the results of the forecast did not enter the confidence interval of the forecast and were lower than the actual values (figure 6). There are many methods for forecasting time series. If the time series is very large, experts recommend dividing it into short segments, each of which should be analysed separately. In the case of our sample of two companies, the range for forecasting was quite large - 11 months. It is possible that the large length of the time series distorted the angle of slope of the regression and for this reason the forecast values differed from the real ones. When studying time series it is very important to pay attention to patterns that are poorly predicted, for this reason it is possible that for our regression analysis we should have taken November samples for the forecast, which would have allowed us to rearrange the trend and make the forecast more accurate. For example, for the Walt Disney stock price in November, an important indicator was the resistance level, which is a price level above which the stock price rarely rises. In our situation, the stock price rose sharply, which was a marker of some event that caused the price to rise upwards. Also, in stock forecasts the focus on fundamental analysis is also necessary - these are the events occurring in the company, news that can dramatically change the value of shares, as well as on the financial indicators of the company. Another way of forecasting data can is moving

11 average, which helps to smooth out price fluctuations and determine the direction of the trend. In our case the exponential moving average (EMA) for both companies may be more suitable than linear regression. Summary The report talks about predicting the future prices of Walt Disney and Netflix stocks using a math method called linear regression, which looks at how stock prices have changed over time to guess their future values. This method uses some statistics to see how strong and accurate these predictions are. For both companies, they tried to guess the stock prices for three months in the future. However, their predictions didn't match very well with the real stock prices, especially for Walt Disney. This means their method of guessing wasn't very reliable for Disney, as the numbers they used to measure how good their prediction was turned out to be low. For Netflix, the prediction was a bit better because the numbers showed a stronger match between their guesses and the actual prices. They used a program called Excel to do these calculations, looking at how stock prices moved over time. In the end, they found that while they could see a general trend in how stock prices were moving, their method of predicting exact future prices wasn't very accurate. This suggests they might need to try different ways of predicting or look at more factors that could affect stock prices to get better at guessing their future values

Related Documents