4
$\begingroup$

I just started learning predictive modelling in R , however i do understand some terms below but i lack in making more interpretations , just want to know what a pro statistician or Data Scientist interpret from it . How does one look at it ?

> summary(model)

Call:
lm(formula = comp$Minutes ~ comp$Units)

Residuals:
    Min      1Q  Median      3Q     Max 
-9.2318 -3.3415 -0.7143  4.7769  7.8033 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept)    4.162      3.355    1.24    0.239    
comp$Units    15.509      0.505   30.71 8.92e-13 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 5.392 on 12 degrees of freedom
Multiple R-squared:  0.9874,    Adjusted R-squared:  0.9864 
F-statistic: 943.2 on 1 and 12 DF,  p-value: 8.916e-13
Improve this question
$\endgroup$
7
  • $\begingroup$ You can just type in R ?summary.lm. The documentation tells you the statistics name of each word, and you can then use wikipedia to learn more about each. $\endgroup$
    – Ricardo Cruz
    Commented Aug 1, 2016 at 13:32
  • 9
    $\begingroup$ Ah, this specific summary! $\beta_0=4.2$ and $\beta_1=15.5$. Your $\beta_1$ is statistically significative (t-test << 0.05). $R^2=0.99$ meaning 0.99 of the variance in the dependent variable is explained by your regression. Finally, the F-test is statistically significative as well. In other words, it looks like a good regression. $\endgroup$
    – Ricardo Cruz
    Commented Aug 1, 2016 at 14:37
  • 1
    $\begingroup$ @RicardoCruz This'd make a good answer :) $\endgroup$
    – Dawny33
    Commented Aug 4, 2016 at 10:52
  • 3
    $\begingroup$ From experience, not from theory, $R^2=.99 $ is an unusually large result. Here is the point: the p-value is calculated under the assumption that the residuals follow a certain pattern, and without checking that you cannot trust the p-value. You have only 14 observations, so a few extreme outliers would easily invalidate the assumptions needed to compute the p-values. Plot the pairs (comp$Minutes, residuals) and check if there are a few large residuals. $\endgroup$
    – VictorZurkowski
    Commented Aug 5, 2016 at 0:23
  • 1
    $\begingroup$ What @VictorZurkowski said. I did not actually "answer" the question because it would involve mentioning all those things. Anyhow, here is a list of the typical assumptions when hypothesis-testing a linear regression. Since you only have one variable, you can plot the regression for that variable and see if the residuals (difference between your line and the points) make sense. The difference should be uniform across the variable. See here how to plot it. $\endgroup$
    – Ricardo Cruz
    Commented Aug 5, 2016 at 10:45

1 Answer 1

Reset to default
1
$\begingroup$

The global test shows that the model is significant as the p value is very low which is p-value: 8.916e-13 in alpha level of 0.05. 98.64% of the variability of Minutes is described by Units. The effect of Units is significant in predicting Minutes. For each one unit value increase in Units, Minutes increases by 15.509.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.