Binary-Tree-for-Large-Dimension-Data-Search
This small preject is try to use binary tree to speedup the data whose dimension is very large.
Develop tools and techniques
- Java
- NetBeans
Step
- Generate 1000000 datas with 10 dimensions, and every dimension is generated by Gaussian distribution
- Construct a binary search tree for every dimension
- Sort query vector from big attribute to small attribute
- Search binary serach tree with accuracy for the biggest dimension of the query and get some similar datas which can be viewed as candidate
- For the remaining candidates, we search other dimension by linear search
- Get a set of data which is similar to the query