0

1) What is the effect of randomSeed parameter on dimensionality reduction by random projection in weka?

2) Secondly it is said that dimensionality reduction does not loss information, But I have observed that if we set the numberOfAttributes smaller, it improves accuracy, Whereas if we set numberOfAttributes close to current(actual) or a large value, whether it reduces accuracy?

Keshava GN
  • 4,195
  • 2
  • 36
  • 47
Kashif Khan
  • 301
  • 6
  • 17

1 Answers1

0

Google for "random number seed" to understand the random seed parameter.

Dimensionality reduction loses information, but the lower dimensionality can make things easier to optimize. The data lost can also be distracting noise, so don't be surprised to see the lossy approach improve performance sometimes. Just don't rely on it.

Has QUIT--Anony-Mousse
  • 76,138
  • 12
  • 138
  • 194
  • it means we are never sure that dimensionality reduction will improve accuracy or not however it will always optimize the classification process ? – Kashif Khan Jan 11 '14 at 15:20
  • No, otherwise we would always or never do it, but not just sometimes. – Has QUIT--Anony-Mousse Jan 11 '14 at 15:31
  • actually i have 146 dimension in my data. I have applied information gain already to reduce the dimension. I want to speed up my classification process without dropping accuracy. I came to know that only way to optimize the classification is to use quick learner and reduce dimensions however in my case they drop accuracy i.e. if i set dimension to reduce attributes to 80% , it reduces accuracy more than i set reduce attributes to 50%, Is there any other way around to optimize the classification process ? – Kashif Khan Jan 12 '14 at 09:18