Abstract
Most academic research methods can be categorized into two types: hypothesis oriented and data oriented. Hypothesis oriented: firstly make a hypothesis or purpose, and prove or disprove it with experiments and theorems. Data oriented: find out unknown principle or property from a great quantity of data. Taking tourism for figure of speech, hypothesis oriented researching is like that after deciding a destination, find out a route to reach there. Data oriented researching is like adventuring, which means that following intricate roads, find out unexpected treasure on the wayside. In this article, the author discusses two mining problems for each type. One is to improve the time complexity of a DNA/RNA sequence mining tool, MotifSeeker. It can find out motifs in DNA/RNA sequences with quite high accuracy. However, generating candidate motifs takes a lot of time because of repeated comparison. Aiming at the bottleneck of running time, we adopted a new strategy. Then prove the improvement in running time with formula deriving and experiment result. The other is to find out the correlation between several databases by observing their content with economical opinion. The achievement is the correlation between the tRNA variables and the codon usage of several organisms.