From 11c73eb18bc62608ad2d8aeb1571ecbaf78e689a Mon Sep 17 00:00:00 2001 From: Okada Haruki Date: Thu, 5 Sep 2019 08:42:11 +0900 Subject: [PATCH] Update README.md --- README.md | 56 ++++++++++++++++++++++++++++--------------------------- 1 file changed, 29 insertions(+), 27 deletions(-) diff --git a/README.md b/README.md index e6a5095..ee59585 100644 --- a/README.md +++ b/README.md @@ -19,33 +19,35 @@ Both HyperLogLog and MinHash require a precision parameter. Basic guidelines are available as follows, and `HLLCounter.MIN_P = 4 <= p <= 18 = HLLCounter.MAX_P`. -####HyperLogLog p @ 99.7% Confidence -p | Relative Error ----:|---: -4 | 75% -5 | 65% -6 | 47% -7 | 32% -8 | 23% -9 | 16% -10 | 10% -11 | 8% -12 | 5% -13 | 4% -14 | 2.5% -15 | 2% -16 | 1.3% -17 | 1% -18 | 0.7% +#### HyperLogLog p @ 99.7% Confidence -####MinHash k @ 99% Confidence -**Relative Error** | **Intersection Size -->** | | | | * -:------------------|--------------------------:|-------:|-----:|------:|-----: -- | 0.01% | 0.1% |1.0% | 5.0% |10.0% -100% | 90000 | 9000 |900 | 170 |75 -50% | 313334 | 31334 |3134 | 587 |280 -25% | - | 116800 |11520 | 2208 |1040 -10% | - | - |68455 | 13128 |6210 +|p | Relative Error| +|---:|---:| +|4 | 75%| +|5 | 65%| +|6 | 47%| +|7 | 32%| +|8 | 23%| +|9 | 16%| +|10 | 10%| +|11 | 8%| +|12 | 5%| +|13 | 4%| +|14 | 2.5%| +|15 | 2%| +|16 | 1.3%| +|17 | 1%| +|18 | 0.7%| + +#### MinHash k @ 99% Confidence + +|**Relative Error** | **Intersection Size -->** | | | | * | +|:------------------|--------------------------:|-------:|-----:|------:|-----:| +|- | 0.01% | 0.1% |1.0% | 5.0% |10.0%| +|100% | 90000 | 9000 |900 | 170 |75| +|50% | 313334 | 31334 |3134 | 587 |280| +|25% | - | 116800 |11520 | 2208 |1040| +|10% | - | - |68455 | 13128 |6210| This MinHash k table can be generated by using `minhash_k.py` in the `utils` directory. For now, the only requirement is scipy, which you can install with @@ -63,4 +65,4 @@ MinHash k: 4800 Error at k: 0.25 ``` -Additional information is available with `./utils/minhash_k.py --help`. \ No newline at end of file +Additional information is available with `./utils/minhash_k.py --help`.