if the value of $R^2$ in linear regression analysis is 1.0, is it good or bad? What are the alternatives?  
The picture shows my results. I have reviewed all the row data many times. Unfortunately, I couldn't find mistakes in data entry. Are there any possible solutions or alternatives in such cases?
 A: In general a better $R^2$ is good (given that you aren't making your model too complex; this is what the adjusted $R^2$ value is for). However, $R^2=1$ means that for some reason your model predicts the response variable perfectly, which is generally too good to be true. Common reasons are:


*

*you have some kind of artificial data (e.g. made up for a homework exercise)

*you are overfitting (using a model that is too complex); this is probably not the current situation, since you have 136 observations and 8 predictor variables

*you messed up something about your data analysis (e.g. you're accidentally predicting something that you meant to be one of the predictor variables)


It's hard to say more without more context about what your variables are and why you're doing the analysis.
A: An $R^2=1$ means that the data is perfectly correlated.  This is reflected in your standard errors being $0$.
When performing linear regression, you want the value of $R^2$ to be as close to $1$ as possible, although there are times that a much lower $R^2$ value can be acceptable.
A: As other answers have mentioned, more background information on your data is needed to draw a reliable conclusion.
With that being said, an R2 value of 1 does mean that your model is potentially "too good to be true" as another answer mentioned.
From your output, I am assuming that your regression is of the Ordinary Least Squares form. I would be inclined to do the following:


*

*Run a Variance Inflation Factor test on your independent variables to test for multicollinearity. This is where two or more independent variables have a strong correlation (and could be theoretically linked). Therefore, the dependent variable is accounting for the effect of this multiple times, but your R2 value is still rising, giving an overly optimistic result. It may be the case that you need to drop certain multi-collinear variables if the theoretical nature (or lack of it) behind them justifies doing so.

*If your data is a time series one, then it is possible that a trend in your data is producing a high R-Squared due to high correlations between the dependent and independent variables, even when this is likely due to the trend which does not have any theoretical foundation. I would be inclined to test your model for autocorrelation (if indeed your data is time series, if not ignore this step) using a Durbin-Watson test, and if this problem is present then try first differencing the values and run the regression again.


Overall, an R2 value of 1 - while possible - indicates perfect collinearity and certainly warrants further investigation before a conclusion can be drawn.
