I do have one for matrices.
"Order of matrix __ is (__x__) and the order of matrix __ is (__x__). Thus ____ is not defined because the number of columns of (First matrix) is not equal to the number of rows in (second matrix)"
You don't actually have to write all of that out.
I'd be succinct and to-the-point and write:
Number of columns of the first matrix (is not equal to) the number of rows of the second matrix.
^^ That would be sufficient for explaining why a product matrix is undefined.
Also, if you were to describe how unique solutions exist, calculate the determinant, and then show det=44(is not equal to 0), and so, unique solutions for the simultaneous equations exist.
Core:
* There is a clear pattern to the non-random scattering of the points. <-- this is what I'd write for a residual analysis of a non-linear relationship scatterplot.
* There is no clear pattern to the random scattering of points <-- residual analysis for a linear relationship.
* Upwards (increasing) trend OR downwards (decreasing) trend. <-- explaining time series trends.
* Seasonal indices:
e.g. ice-cream sales 1.7 in summer
Ice cream sales in summer are 70% above the average seasonal ice cream sales for 2012.
When describing a relationship between two numerical scatterplots in terms of direction, form and strength, be straightforward and just say: strong, positive, linear relationship. Don't waffle on.
When describing which is the best data transformation (for negative relationship)
~ This transformation yields an r value closest to -1.
~ This transformation yields an r
2 value closest to 1.
When describing which is the best data transformation (for positive relationship)
~ This transformation yields an r value closest to 1.
~ This transformation yields an r
2 value closest to 1.
Your intercept, slope & coefficient of determination written explanations are fine