"
Team:Davidson-Missouri Western/Results
From 2008.igem.org
Krmuscalino (Talk | contribs) |
Krmuscalino (Talk | contribs) |
||
| (3 intermediate revisions not shown) | |||
| Line 10: | Line 10: | ||
'''Minimal Collision Test and Predictability Test''' | '''Minimal Collision Test and Predictability Test''' | ||
| + | |||
| Line 20: | Line 21: | ||
Note: Twist model’s graph results vary depending on constraints (all other models do not). | Note: Twist model’s graph results vary depending on constraints (all other models do not). | ||
Depending on constraints Twist is better than Net and Spoke, but Split is always equal or better in performance. | Depending on constraints Twist is better than Net and Spoke, but Split is always equal or better in performance. | ||
| - | + | ||
| + | |||
'''Unequal Probabilities''' When random generated messages based on unequal probabilities of 128 ascii characters occurring. | '''Unequal Probabilities''' When random generated messages based on unequal probabilities of 128 ascii characters occurring. | ||
| + | |||
'''1) Twist 2)Split 3) Net 4) Spoke''' | '''1) Twist 2)Split 3) Net 4) Spoke''' | ||
Latest revision as of 22:12, 29 October 2008
Ranking of Models (1 – best; 4 – worst)
Primary Property Test - All models were superior.
Minimal Collision Test and Predictability Test
Equal Probablities When random generated messages are based on all 128 ascii characters occurring at equal probabilities.
1)Split 2)Twist 3) Net 4) Spoke
(Evaluation based on patterns and white space in scatter plots).
Note: Twist model’s graph results vary depending on constraints (all other models do not). Depending on constraints Twist is better than Net and Spoke, but Split is always equal or better in performance.
Unequal Probabilities When random generated messages based on unequal probabilities of 128 ascii characters occurring.
1) Twist 2)Split 3) Net 4) Spoke
Note: Twist model’s graph results vary depending on constraints (all other models do not). Depending on constraints Twist is better than Split.
Overall analysis indicates that the more all characters interact with each other under a hash function, the more ideal the hash function.
