Regression Analysis and Hypotheses - Research Paper Example

?Regression Analysis: A mathematical tool that is used in case of linear operations for comparisons and evaluation of two or more variables. It relates to the statistic branch of mathematics. It is also defined as the usage of data for yielding of a relationship between number of parameters and variables that are dependent and independent (Rae, 2008). The final outcome can be in form of plot, an equation derived or the net value. The effect so determined is in form of casual relationship between two variables, where the upward shoot or downward shoot in the value directly affects the other variable. Regression analysis finds large number of principles and applications in economics. It is an effective tool of mathematical analysis, which allows estimation and evaluation to the nearest proximity compared to the past. It also allows predicting a certain future value through comparison and combination of two or more variables. Usually, hypotheses are taken up for case estimation and evaluation which involve plotting of various variables dependent on the time factor. Regression models can be both linear and non linear in nature. Response variables are generated by the comparison of numerous variables that are involved in the computational process. Response variable is of quantitative nature and derived as result of other variables tested and compared amongst each other. Regression analysis is used for predicting certain event (mostly the dependent variable), its outcome in terms of numerical percentage and value in combination with other variables (independent) which might be elements of the present time (Seber & Lee, 2012). Equations are derived which contain both the dependent and independent variables, against some constant. Regression analysis equations are linear in nature. The function so obtained is a straight line, in form of linear equation. Often the natures of the dependent variables have direct impact on the kind of final linear equation being yielded. In case of the dependent variable being dichotomous, the logistic form of regression is handy in such cases for evaluation of the linear plots. When multiple independent variables are up against one dependent variable, it constitutes standard linear multiple regression (Hill & Lewicki, 2006). In such cases, the single dependent variable can be compared one on one against the respective independent variable. Numbers of variables are set against each other and each of them affects the other directly. For example in case of stock exchange the stability of and sound running of stocks are dependent on various external factors, which can include the economic condition, the security and safety, any major political decisions in the governmental corridors. Any unexpected decisions are ascribed to the rise or fall in the stocks value. In this case, the stock market is one variable which is dependent on number of variables that are aforementioned. Each of these variables, both dependent and independent are plotted against a graph with x-axis representation and y-axis along a straight linear line. Stocks standing = Stable Market, Political decisions, other variables plus a Constant. For any economic factor indication, the variables that can be taken in to account include the net flow of income at a particular instance of time, the dependent variables on the fixed amount, the surplus amount in form of the revenue generated and the possible depreciation of the capital. All these are equated and compared against each other in form of plot to estimate the future trends. Mostly, it involves number of variables being operated upon; however at a particular instance of time when one variable is under observation, it constitutes a simple regression analysis scenario. For example in case of an individual’s average spending against his profession, his resources, his requirements, all these alone make for attributes which are evaluated and analyzed for a certain desired future scenario through the regression tool of estimation. If for a given scenario, more number of variables are taken into an account for comparison, such a scenario results in multiple regression. It is handy and applicable only when numbers of variables are impacting a single unit directly at some instance of time. Dependent and independent variables can be also plot against each other in form of scatter plot function where each entity corresponds to other on a particular value. Each point is representative of a particular value. For accurate results that may be desired, the factors of consideration include, the accuracy of the data (both the dependent variables and independent), equal and accurate weighing of the number of cases (in dependent variables). Histogram analysis can be used for determining the normality and smoothness of curve. Histogram would help identifying the pattern obtained and the results yielded. Any variable that results in uneven shoot of the data is highlighted to the histogram tool. It can be used for any variable or can show the entire result as the end product. Other kinds of plots that are possible include the probability plots, which are more loosely arranged based on the assumption and estimation. In case of probabilistic graphs, each variable is directly in association with other variable along the opposite axis (Billard & Diday, 2012). Finally, the normality can be judged through the process of residual evaluation. It is the process in which the estimated figures are evaluated and compared against the practical values that exist. Normality reading gives an idea of the degree of accuracy of the analysis and estimation process conducted (Cameron & Trivedi, 1998). The scattered pattern so obtained can be largely due to the difference between the values predicted and actual values; this all is possible through the normality findings procedure. Based on the differences between the values obtained, and the ground values, the differences can be reduced through appropriate methods. The normality curve and value is sort of pilot test, which gives a rough idea of the situation at hand and helps correcting the findings in due time. Hence regression analysis is an important tool widely being used in market research and large businesses and economic factors heavily make use of this technique. Works Cited: Billard, L., & Diday, E. (2012). Symbolic Data Analysis: Conceptual Statistics and Data Mining. John Wiley & Sons.Print. Cameron, A. C., & Trivedi, P. K. (1998). Regression Analysis of Count Data. Cambridge University Press. Hill, T., & Lewicki, P. (2006). Statistics: Methods and Applications : a Comprehensive Reference for Science, Industry, and Data Mining. StatSoft, Inc.Print. Rae, A. (2008). Organizational Behavior: An Introduction To Your Life In Organizations. Pearson Education.Press. Seber, G. A., & Lee, A. J. (2012). Linear Regression Analysis. John Wiley & Sons. Read More