REGRESSION ANALYSIS
Introduction
Regression analysis is a statistic that is used to determine the relationship between the variable that is the independent variable (predictor) and the dependent variable (result). It used to predict the future outcomes by studying the effects of the independent variables on the dependent variable, for instance, a farmer can use regression analysis to predict the agricultural yield by examining the pattern of rainfall amount. In this case, the return is the dependent variable, and the amount of rainfall is the independent variable. The amount of rain will predict the yield. We expect to have more yields when the moisture is enough because there is a relationship between the amount of rainfall and the return. However, other variables may affect the yield but are much insignificant. Such factors are referred to as error terms. To determine the relationship between the variables under consideration, we a straight line that will also help us understand the slope.
Application of regression analysis
Regression analysis is widely applied in business organizations and generally the economic sector. Business organization are more interested in making a profit in the short run or long run and therefore will use regression analysis to study and analyze the factor that will affect the profitability of the business so that they can do control and forecast the output. The company will minimize the insignificant factors, so their impact is very negligible. For instance, the sale may be predicted by a different factor such as customer satisfaction, weather conditions and so on so forth. Statistics will formulate the hypothesis that the level of customer satisfaction has an impact on sales. After the formulation of the theory, regression analysis is done to ascertain if the level of customer satisfaction can be used to predict the amounts of transactions. Generally, regression analysis will help in controlling the variables that influence the number of sales and as well as forecast future sales
Advantages of Regression analysis
Regression analysis is simple to understand and interpret because it can be presented graphically. By graphically drawing the regression line, we easily deduce the conclusion by looking on how the variable is distributed. The regression line drawn gives the regression model that can be used to predict future outcomes
Regression analysis can be carried out on statistical software such as SPSS and STATA. Ultimate, it becomes easier to do calculations
Disadvantages of regression analysis.
When carrying out regression analysis hypothesis are formulated. However, we may, for instance, accept the null hypothesis while it should be admitted that is committing type I error or take the alternative hypothesis while it is true leading to type 11 error. Consequently, this may result in making the wrong decision that may imply the forecasted outcome. For example, the hypothesis may be formulated that female vice is higher Hz than the male voice. The theory may not be entirely correct because under some circumstances such as males are nervous their sound may higher than that of females and hence it is a limitation.
Linear regression assumes that data are independent. For instance, the marks one person has nothing to with marks of another person. However, this is not the case because students in the same class tend to be similar and they are taught with the same teacher.
Linear regression analysis only looks at the mean of the dependent variable, and therefore it is not a complete description of the relationship of variables.
Problem
Regression analysis plays a significant especially in the business sector since it allows firms and managers to make informed decisions that will enhance the progress and growth of the businesses at large. However, considering the limitations of this statistical technique, there is a need for adjustments. The regression model may have negative implications when the hypotheses formulated are not entirely correct as this to making wrong decisions. Regression analysis is sensitive to outliers. The outliers are another factor that may influence the dependent variables but are not under the study. Since the mean of an outlier is assumed to be zero, it causes discrepancies due to the assumption, and this may be of good when making a prediction. In spite of the premise underlying analysis, it has been applied in diversified economic sectors. If the outlier be taken into consideration and they’re perfect mean taken into account, then a more composed regression model would arrive at that would eliminate the assumptions.
Analysis
The analysis is done through the formulated hypothesis which could state that regression analysis is not a perfect measure for the relationship between variables to be used in forecasting future outcomes. However, basing our on the advantages and disadvantages of regression analysis, it is evident if statistical adjustment is made to the model then the mean of the error term would not be zero. This will take care of residuals. Once this is done, we will then regression would offer a perfect relationship between variables.
Critique
Regression analysis has weakness since the correlation between variables cannot correctly predict the future result. Due to this, it becomes inefficient in modeling linear situations in which variables are related. Since outliers are spread uniformly along the line of best fit, we end up having the mean of the error term being zero. As a result, we cannot correctly conclude that the independent variables under consideration are the only one that determines the dependent variable. Linear regression, on the other hand, does not give the standard error and we may have type one mistake and type two error when the null and alternative hypothesis are misinterpreted that is accepting the null hypothesis while it should be rejected and rejecting the alternative hypothesis while it should be allowed.
Conclusion
In conclusion, regression analysis plays a profound role in the economic sector by using mathematics as a tool to predict sales patterns in the market. Most managers make decisions based on the outcome of the linear regression analysis. The regression model generated will provide advice to the executive on the future market patterns and issues when incorporating the factors that determine the outcome of the dependent variables.
References