Thursday, November 12, 2020

Logarithms

Logarithms

The following is a nice way to think about logarithms:
logba\log_b a \equiv The number of times aa is divided into bb parts such that each part becomes 11.

For example, 10001000 needs to be divided into 10 parts 33 times to get to 1, so log101000=3\log_{10}1000 = 3.

Reaching arbitrary numbers by repeated division

Now, we may ask—how many times should aa be divided into bb parts to reach an arbitrary number xx? We can split aa reaching 11 to aa reaching xx and then xx reaching 11.

Let kk \equiv The number of times aa is divided into bb parts to reach xx.

We have,
logba=k+logbx    k=logbalogbx=logb(ax) \log_b a = k + \log_b x \\ \implies k = \log_b a - \log_b x = \log_b \left (\frac a x \right )

Intuition for the change of base rule

We will use this to build intuition for the following identity:
logqa=logpa.logqp \log_qa = \log_pa . \log_qp

Informally, the above asks,
how do we relate the number of times we divide aa in qq parts to reach 11 and the number of times we divide aa in pp parts to reach 11?

Let us look at the first step of calculating logap\log_a p. We divide aa into pp parts, and each part is of size a/pa/p. Just to perform this first step, if I was only allowed to divide by qq, how many divisions would I need to reach a/pa/p ?

Well, this is the question we asked in the previous section!
The number of divisions is equal to logqaa/p=logqp\log_q \frac{a}{a/p} = \log_q p. Notice this is independent of aa, so this holds true to reach any x/px/p from xx, in other words, it holds for every step of the calculation of logap\log_a p.

So, we have logqp\log_q p divisions [in the process where you only divide by qq to reach 11] for each of the logpa\log_p a divisions [in the process where you only divide by pp to reach 11].

Which gives us:
logqa=logpa.logqp\log_q a = \log_p a. \log_q p

Yayyy!

I would love to be able to formalize the intuition I have here, and also extend it for other logarithm identities. If you have questions, or anything that may help, feel free to comment, or reach out to me at mahathivempatiresearch@gmail.com.

Share:

0 comments:

Post a Comment