Re: [SPATIAL] Best Fit Calculation

Thanks. I added my comment on the issue. I think we should revert and then someone can put up a patch to make this pluggable. As it stands, this Best Fit calculation has nothing to do with the CartesianTierPlotter anyway, so we could refactor it pretty easily. -Grant On Apr 14, 2010, at 12:

Re: [SPATIAL] Best Fit Calculation

On Apr 14, 2010, at 12:12 PM, Chris Male wrote: > Hi, > > On Wed, Apr 14, 2010 at 6:07 PM, Grant Ingersoll wrote: > > On Apr 14, 2010, at 11:06 AM, Chris Male wrote: > > > Hi, > > > > My understanding of the benefits of the new algorithm is that it means a > > lower tier level resulting in f

Re: [SPATIAL] Best Fit Calculation

Tables are well on JIRA : https://issues.apache.org/jira/browse/LUCENE-2359 Nicolas 2010/4/14 Helleringer, Nicolas > Here are the summary tables : > > First a table to remind metrics on the Tiers : > Tile Level TierLegnth TierBoxes TileLength (

Re: [SPATIAL] Best Fit Calculation

Here are the summary tables : First a table to remind metrics on the Tiers : Tile Level TierLegnth TierBoxes TileLength (miles) 0 1 1 24902 1 2 4 12451 2 4 16 6225,5 3 8 64 3112,75 4 16 256 1556,375 5 32 1024 778,1875 6 64 4096 389,09375 7 128 16384 194,546875 8 256 65536 97,2734375 9 512 262144 4

Re: [SPATIAL] Best Fit Calculation

On Wed, Apr 14, 2010 at 6:24 PM, Yonik Seeley wrote: > On Wed, Apr 14, 2010 at 12:12 PM, Chris Male wrote: > > On Wed, Apr 14, 2010 at 6:07 PM, Grant Ingersoll > >> On Apr 14, 2010, at 11:06 AM, Chris Male wrote: > >> > For those doing just Cartesian Tier filtering it seems like the new > >> > a

Re: [SPATIAL] Best Fit Calculation

On Wed, Apr 14, 2010 at 12:12 PM, Chris Male wrote: > On Wed, Apr 14, 2010 at 6:07 PM, Grant Ingersoll >> On Apr 14, 2010, at 11:06 AM, Chris Male wrote: >> > For those doing just Cartesian Tier filtering it seems like the new >> > approach is a win, but for those doing distance calculations on t

Re: [SPATIAL] Best Fit Calculation

Hi, On Wed, Apr 14, 2010 at 6:07 PM, Grant Ingersoll wrote: > > On Apr 14, 2010, at 11:06 AM, Chris Male wrote: > > > Hi, > > > > My understanding of the benefits of the new algorithm is that it means a > lower tier level resulting in fewer boxes, but more documents inside those > boxes that are

Re: [SPATIAL] Best Fit Calculation

On Apr 14, 2010, at 11:28 AM, Helleringer, Nicolas wrote: > That minTile param allows you to trade off between filtering accuracy > and faster tile filtering. Without the param (or until it can be > implemented) the correct approach seems like the above, without a > minTile. This sounds to me l

Re: [SPATIAL] Best Fit Calculation

On Apr 14, 2010, at 11:06 AM, Chris Male wrote: > Hi, > > My understanding of the benefits of the new algorithm is that it means a > lower tier level resulting in fewer boxes, but more documents inside those > boxes that are outside of the search radius. > > While having fewer boxes means few

Re: [SPATIAL] Best Fit Calculation

> > That minTile param allows you to trade off between filtering accuracy > and faster tile filtering. Without the param (or until it can be > implemented) the correct approach seems like the above, without a > minTile. This sounds to me like the old approach is correct. > minTier and maxTier at

Re: [SPATIAL] Best Fit Calculation

I ll try to find a little bit of time tonight to make a sample data set go through the two calculations to see the differences. I ll make a summary table. I ll comment the issue with some comments on 'my' version of the alogrithm right now. Nicolas 2010/4/14 Chris Male > Hi, > > My understandi

Re: [SPATIAL] Best Fit Calculation

On Wed, Apr 14, 2010 at 11:06 AM, Chris Male wrote: > While having fewer boxes means fewer term queries to make against the index, > more documents means more costly calculations to filter out those extraneous > documents. Filtering out documents (greater selectivity) seems like it should be the

Re: [SPATIAL] Best Fit Calculation

Hi, My understanding of the benefits of the new algorithm is that it means a lower tier level resulting in fewer boxes, but more documents inside those boxes that are outside of the search radius. While having fewer boxes means fewer term queries to make against the index, more documents means mo

[SPATIAL] Best Fit Calculation

LUCENE-2359 changed the best fit calculation. I admit, I'm not entirely certain which one is right, so I thought we should step back and talk about what we are trying to achieve. Please correct me if/where I am wrong. Looking at the problem of tiers/tiles/grids in general, we are taking a sphe