Data Analysis by Driven IFS

Sample

e can use the Address Stats data to begin to assess how the likelihood of landing in a bin is affected by the previous bin occupied. For example, if there were no relation between the current bin and the next, then

p_i*p_j = p_ij

where pi is th probability of landing in the ith bin, and

pij is the probability of landing in the ith bin immediately after landing in the jth bin.

In this example, using the 1000 data points we calculate

p3 = 0.184	p4 = 0.345
p1 = 0.315	p2 = 0.156

Some dependences are obvious:

p₄*p₄ = 0.119 yet we observe no points with address 44.

Slightly more subtly,

p₄*p₁ = p₁*p₄ = 0.109 yet we observe p₁₄ = 0.177 and p₄₁ = 0.005.

p₄*p₃ = p₃*p₄ = 0.063 yet we observe p₃₄ = 0.086 and p₄₃ = 0.184.

We note p₄₃ = p₃ because every data point in bin 3 is followed immediately by a data point in bin 4: the sqaures with addresses 13, 23, and 33 are empty.

Return to Equal-Size Bin Sample.