Business Statistics And Forecasting Assignment Sample

• Type Assignment
• Downloads840
• Pages8

Introduction

Get free samples written by our Top-Notch subject experts for taking online Assignment Help services.

The link among a group of independent variables as well as a dependent variable is formally described by the "regression analysis". There are many multiple kinds of regression equations which could be used. Such decision is frequently influenced by the kind of information available for the dependent variable as well as the kind of strategy which delivers the greatest fit. In this report, it would go through the most popular kinds of regression analysis as well as how to choose the best one for particular dataset. It would offer everyone an outline as well as facts on the way to assist users decide. It categorizes regression methods depending on the various categories of independent variables and one dependent variable. When people are not sure whatever approach should employ, then figure out what kind of dependent variable that have along with then concentrate by the side of that area of this text. Following procedure might aid in narrowing down the options. It would go through "regression models" which work well with ongoing, categorized, as well as count data as the source of the dependent variable.

Methodology

The "multiple regression strategy" is being used for analyzing the correlation among the dependent as well as multiple independent variables (Kadim and Nardi, 2018). The method is depicted through the formula Y is equivalent to such an addition of bX1, cX2, dX3 as well as E, in which Y is the "dependent variable", on the other hand X1, X2, X3 are "independent variables". The slopes of the formula are indicated by the b, c & d while the residual value is represented by the E. The equation is given below:

y = mx1 + mx2+ mx3+ b

where;

Y indicates the "regression's dependent variable"

M denotes the "regression slope."

x1 represents the "regression's initial independent variable"

x2 indicates the "regression's next independent variable"

x3 denotes the "regression's last independent variable"

B is a constant.

The data set which is used here is related to the stock market. The data set is collected through the given in data sources. As a data source it is chosen by Yahoo Finance. The data set which is selected for this project is of the BP PLC company. The data set which is downloaded from the online source is as csv format (Agung, 2018). Then in the next stage the csv file is imported to the EViews software. After importing the csv file, the multiple regression analysis, Descriptive analysis, time series analysis and forecasting is performed.  For performing the analysis one dependent variable and three independent variables is selected. For dependent variables it is chosen the volume of the BP Plc as well as three independent variables are open, high and low.

Descriptive statistical analysis is being used for characterizing the fundamental characteristics of data in research (Bensalma, 2021). They give concise precises of the example as well as measurements. They are the foundation of almost any quantitative analysis, along with rudimentary data presentation. In most cases, the "descriptive statistics" are differentiated through "inferential statistics". The "Descriptive statistics" actually describe what is and what the information reveals. Using the "inferential statistics", it is attempting on the way to draw inferences that go beyond the current information. For example, it utilizes inferential analysis to attempt to infer whatever the public would believe based on the sample information (Indriaty et al. 2020). Alternatively, it utilizes inferential analysis in the direction of assessing the likelihood how a reported distinction is reliable or that this occurred via chance throughout this research. Thus, it utilizes "inferential analysis" on the way to derive generalizations using the data, whereas "descriptive statistics" just explain what's going on in the information. [Refer to appendix 1]

The goal of the "descriptive analysis" is to characterize or summarize information. Although this does not anticipate the future, this may nonetheless be incredibly useful in commercial settings. It is primarily due to the fact that "descriptive analysis" creates information easier in the direction of consuming, that also creates it simpler for experts to behave on (Zhao, 2019). Additional advantage of the "descriptive analysis" is that this could aid in the removal of irrelevant data. This is because the numerical methods employed in this sort of study typically focus on data trends rather than outliers. The "descriptive analysis" technique generally consists of the identical limited phases.

Collect data

The basic stage in whatsoever sort of data analysis is to gather the information. This could be achieved in a diversity of methods, although good old-fashioned observations are frequently employed.

Clean data

Cleaning the information is a crucial stage in descriptive as well as other forms of data analysis. It is because information might be structured in unreadable ways, making data impossible for interpret. Cleansing information can entail converting this to a different textual form, classifying information, and/or removing abnormalities.

Apply methods

Ultimately, descriptive analysis entails using the selected statistical methodology to get the required results. What methodologies it uses would focus upon the information it is working with & what it is trying to figure out. 

Finding & discussion

This above data set is obtained from the least squares. The R squared value, adjusted R squared value, regression value, sum squared value, log value, f statistic value probability value is obtained from the least square. Also, some other values can be seen in the table. [Refer to appendix 2]

The above graphical representation illustrates the bar plots which are obtained from the regression analysis. From the bar plot the value of Mean, Median, maximum value, minimum value, standard deviation value can be known. [Refer to appendix 3]

While it has a small set of data based upon the quantity of the independent variables, and whenever the independent variables are strongly linked, "partial least squares" (PLS) regression is beneficial. PLS, like Principal Component Evaluation, reduces the independent variables towards a smaller proportion of statistically independent elements (Agung, 2021). The technique subsequently uses such parts instead of the actual information to conduct regression analysis. PLS focuses on constructing prediction models rather than filtering characteristics. In contrast to OLS, users could incorporate a number of dependent variable factors. The correlation architecture is used by PLS in the direction of finding smaller impacts as well as describe multidimensional trends in the dependent variable. [Refer to appendix 4]

A brief scan of the findings indicates that the variables are statically important as well as the fitting is quite perfect. Nevertheless, if indeed the standard error is sequentially linked, the calculated OLS error terms are incorrect, as well as the correlation parameters are biased as well as unreliable attributed to the prevalence of a delay just on correct side (Aljandali and Tatahi, 2018). As there is an unobserved heterogeneity variable upon that main scale, the Durbin-Watson estimate is inapplicable as a check of "serial correlation" non particular scenario.

For generic, elevated ARMA faults, the "serial correlation test" performs the "Breusch-Godfrey Lagrange multiplier test". Specify the ultimate example of time series to be examined inside the Lag Requirement dialogue box. [Refer to appendix 5]

The "Breusch-Pagan-Godfrey Test" (also known as the "Breusch-Pagan Test") is an experiment for homoscedasticity in logistic failures. Heteroscedasticity implies "differently distributed," in contrast to homoscedasticity, that indicates "same dispersion." Heteroskedasticity is a crucial condition in regression; if this has been rejected, one will be unable to conduct multiple regression.

Forecasting

The "standard deviation" of the "forecast distributions" is about the similar as the mean difference of a residuals while predicting one step forward (Zulfikar and STp, 2019). In reality, if no variables are to be evaluated, as seems to be the situation with the naive technique, the 2 "standard deviations" are equal. For forecasting techniques incorporating parameter estimates, the prediction distribution's standard error is significantly bigger than that of the residue standard error, however this discrepancy is frequently overlooked. [Refer to appendix 6]

Forecasting is an essential aspect of the job of a marketing executive or a stock market analysis. Forecasting sales is crucial to understanding marketplace segment as well as competitiveness, future growth requirements, including selling factors like as incentives, price, marketing, as well as delivery.

Conclusion and recommendation

The aim of creating a "time series model" is almost similar as the purpose of creating various kinds of forecasting analytics for producing an arrangement with the smallest feasible error among the anticipated value of the target attribute as well as the real value. The main distinction among the "time series models" as well as other kinds of arrangements is that "time series models" utilize lag value systems of a target attribute as response variable, so although traditional forms are using other factors as predictor variables, as well as the notion of a lag value does not pertain so because findings do not depict a chronology. After performing the experiment, it brings some recommendations. These are as follows: To begin, create basic designs. Using a large number of independent variables does not always imply that the strategy is excellent. The next stage is to try to generate as numerous regression models as possible using various variable combinations. Then it may put every one of these prototypes together to make an ensemble. This could assist users come up with a good concept.

References

Journals

A Kadim, K. and Nardi, S., 2018. Eviews Analysis: Determinant Of Leverage And Company’s Performance. Global and Stochastic Analysis (GSA), 5(7), pp.249-260.

Agung, I.G.N., 2018. Advanced Time Series Data Analysis: Forecasting Using EViews. John Wiley & Sons.

Agung, I.G.N., 2021. Quantile Regression: Applications on Experimental and Cross Section Data Using EViews. John Wiley & Sons.

Aljandali, A. and Tatahi, M., 2018. Economic and Financial Modelling with EViews. A Guide for Students and Professionals. Switzerland: Springer International Publishing.

Bensalma, A., 2021. An Eviews program to perform the fractional Dickey-Fuller test.a

Startz, R., 2019. EViews Illustrated. University of California: Santa Barbara, CA, USA.

Indriaty, L., Thomas, G.N., Hidayati, N. and Suryati, L., 2020. EVIEWS ANALYSIS: MODEL OF INVESTMENT OPPORTUNITY SET (IOS) AND ITS IMPLICATION TO CORPORATE VALUE ON MANUFACTURING COMPANIES LISTED IN INDONESIA STOCK EXCHANGE (IDX). Palarch’s Journal Of Archaeology Of Egypt/Egyptology, 18(1), pp.366-378.

Zhao, Y., 2019, January. Analyzing the Influence of China's Import and Export Trade on Economic Growth Based on Eviews. In 2018 6th International Education, Economics, Social Science, Arts, Sports and Management Engineering Conference (IEESASM 2018) (pp. 45-49). Atlantis Press.

Zulfikar, R. and STp, M.M., 2019. Estimation model and selection method of panel data regression: an overview of common effect, fixed effect, and random effect model.