Chat with our AI personalities
Yes. The data must not be allowed to influence the choice of the hypothesis.
First you must decide what is a "suitable degree of accuracy" for a particular problem. In many cases, 4 or 5 significant digits are appropriate, or even 3. But it depends a lot on the original data (the final result is not supposed to look more accurate than the accuracy you can justify from the original data), and the purpose of the data (in some cases you need a higher accuracy than in others).
no * * * * * Yes, almost always. If you have n data points which are 1-to-1, then it is always possible to fit a polynomial of degree n-1 or greater.
1. Discovery or identify trends, particularly in time related data 2. Compare sets of data and identify relationships 3. Identifying points that may be erroneous because they are outside of the normal grouping of data. 4. Examine degree of consistency or scattering of data 5. Graphs can effectively communicate ideas/ relationships to others. People can see relationships easier than just looking at numbers.
Interval Data: Temperature, Dates (data that has has an arbitrary zero) Ratio Data: Height, Weight, Age, Length (data that has an absolute zero) Nominal Data: Male, Female, Race, Political Party (categorical data that cannot be ranked) Ordinal Data: Degree of Satisfaction at Restaurant (data that can be ranked)