 Excuse me, on a basis of comprehensive discussion on seemingly uncorrelated regression analysis, अग93 B can easily decide that on the basis of data, on the basis of theory! on the basis of literature review, on the basis of empirical justification, on the bages of methodological discussion, We can conclude this that when we apply seemingly uncorrelated regression, आप से पहले आप थुरेटिकली दिसकस करेंगे, वैदर दा इशु like can be addressed by applying the OLS or by applying the SOAR analysis, number one. Number two, there are some theoretical background. You have some literature review, जो बताता है के खेरिज क्या कैती हैं कि वैदर इस सिच्वोशन में, यहांपार इंके error correlate कार सकते है कि नहीं? अगर वो कैती है के yes, they can. हम सीमिगली on correlated regression apply कार सकते हैं. अगर literature कैता है कि नहीं, इस सिच्वोशन में error कभी भी correlate नहीं कार सकते है, तें हम SOAR को युज नहीं करेंगे, हम आगे मुप कार नहींगे. The next step is to collect the data on the relevant variable. After collection of the data, you have statistical software. And even a STATA is helpful in estimation of SOAR analysis. Or on the basis of SOAR analysis, you can simply find the unit to test. You can simply check the variance inflating factor in terms of VIF for multicollinearity. Then you can simply check the correlation coefficient. In cross-sectional analysis, you can check whether they are correlated or not correlated. If errors are correlated, then you have to discuss that yes, the best approach is SOAR analysis. So on the basis of this, we will apply a whole procedure in the same pattern. If you have time series data, you have to check the stationarity to check whether there is a problem of unit root or there is not a problem of unit root. And on the basis of unit root analysis, you have to check whether there is long run association or there is short run association. If there is long run and short run association, then what is the speed of adjustment? What is happening with the passage of time? Whether they are converting towards equilibrium or they are diverting from equilibrium. Then at the end, you have to give policy recommendations on the basis of your finding, on the basis of your empirical work. So empirical work, if you have an authentic work, you have chosen relevant variable. You have chosen relevant model. You have added relevant functional forms. And on the basis of that, you will give recommendations that will also be relevant. And when you intervene in that market, then that will be result oriented intervention. But if there is a problem in your functional form, there is a problem in your data, you are irrelevant model. As I said, if you have the data pattern inverted U and you apply a simple linear model on it, then it will not minimize your errors. Your findings will not make them efficient. The results will not be policy relevant. They can be just a research paper or a research study. But in the market, intervention will not be possible for that.