X-Git-Url: https://git.njae.me.uk/?a=blobdiff_plain;f=hillclimbing-results%2Fhillclimbing-results.ipynb;fp=hillclimbing-results%2Fhillclimbing-results.ipynb;h=2927601131b394d32d49246625ce7f362d955093;hb=8777c2615919337fda181c43f1949d0edb522ddd;hp=3877ce68184a88729ffab64c973c7de5b43234e4;hpb=c78432a60064579208e363a1a08f930d7c1e326d;p=cipher-tools.git
diff --git a/hillclimbing-results/hillclimbing-results.ipynb b/hillclimbing-results/hillclimbing-results.ipynb
index 3877ce6..2927601 100644
--- a/hillclimbing-results/hillclimbing-results.ipynb
+++ b/hillclimbing-results/hillclimbing-results.ipynb
@@ -2,12 +2,13 @@
"cells": [
{
"cell_type": "code",
- "execution_count": 2,
+ "execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"from datetime import datetime\n",
"import glob\n",
+ "import math\n",
"import pandas as pd\n",
"import csv\n",
"import matplotlib as mpl\n",
@@ -17,450 +18,158 @@
},
{
"cell_type": "code",
- "execution_count": 2,
+ "execution_count": 16,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
- "-5439.653663160256 -8354.182366165229 sa-random-unigram-uniform.csv\n",
- "-5439.653663160256 -8259.44168109899 hillclimbing-random-unigram-uniform.csv\n"
+ "-3935.561885011543 -2494.549133086381 sa-random-unigram-uniform.csv\n",
+ "-3862.1721721032586 -2494.549133086381 hillclimbing-random-unigram-uniform.csv\n"
]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "(-4000, -2400)"
+ ]
+ },
+ "execution_count": 16,
+ "metadata": {},
+ "output_type": "execute_result"
}
],
"source": [
+ "unigram_ylim = None\n",
"for f in glob.glob(\"*unigram*.csv\"):\n",
" df = pd.read_csv(f)\n",
- " print(df.fitness.max(), df.fitness.min(), f)"
+ " y1 = df.fitness.max()\n",
+ " y0 = df.fitness.min()\n",
+ " print(y0, y1, f)\n",
+ " if unigram_ylim:\n",
+ " unigram_ylim = (min(unigram_ylim[0], y0), max(unigram_ylim[1], y1))\n",
+ " else:\n",
+ " unigram_ylim = (y0, y1)\n",
+ "unigram_ylim = (math.floor(unigram_ylim[0] / 200) * 200, math.ceil(unigram_ylim[1] / 200) * 200)\n",
+ "unigram_ylim "
]
},
{
"cell_type": "code",
- "execution_count": 3,
+ "execution_count": 14,
"metadata": {},
"outputs": [
{
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "-14681.308607565503 -27211.09615617547 hillclimbing-random-trigram-uniform.csv\n",
- "-14681.308607565503 -17464.568516864027 hillclimbing-given-trigram-uniform.csv\n",
- "-14681.308607565503 -21515.898852481398 sa-given-trigram-gaussian.csv\n",
- "-14681.308607565503 -17464.568516864027 hillclimbing-given-trigram-gaussian.csv\n",
- "-14681.308607565503 -28346.7456787418 sa-random-trigram-uniform.csv\n",
- "-14681.308607565503 -21065.204759662218 sa-given-trigram-uniform.csv\n"
- ]
+ "data": {
+ "text/plain": [
+ "-3800"
+ ]
+ },
+ "execution_count": 14,
+ "metadata": {},
+ "output_type": "execute_result"
}
],
"source": [
- "for f in glob.glob(\"*trigram*.csv\"):\n",
- " df = pd.read_csv(f)\n",
- " print(df.fitness.max(), df.fitness.min(), f)"
+ "math.ceil(-3935.561885011543 / 200) * 200"
]
},
{
"cell_type": "code",
- "execution_count": 6,
- "metadata": {
- "scrolled": true
- },
+ "execution_count": 17,
+ "metadata": {},
"outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "-12382.205332762649 -6762.002908994095 hillclimbing-random-trigram-uniform.csv\n",
+ "-8523.074815108963 -6762.002908994095 sa-given-trigram-uniform-50.csv\n",
+ "-12897.908909587839 -6762.002908994095 sa-random-trigram-uniform-50.csv\n",
+ "-8347.978847763903 -6762.002908994095 hillclimbing-given-trigram-uniform.csv\n",
+ "-11531.53274337052 -6762.002908994095 sa-given-trigram-gaussian.csv\n",
+ "-8347.978847763903 -6762.002908994095 hillclimbing-given-trigram-gaussian.csv\n",
+ "-12999.784533122312 -6762.002908994095 sa-random-trigram-uniform.csv\n",
+ "-8612.777635758275 -6762.002908994095 sa-given-trigram-gaussian-50.csv\n",
+ "-10935.574066404368 -6762.002908994095 sa-given-trigram-uniform.csv\n"
+ ]
+ },
{
"data": {
- "text/html": [
- "
\n",
- "\n",
- "
\n",
- " \n",
- " \n",
- " | \n",
- " | \n",
- " fitness | \n",
- "
\n",
- " \n",
- " worker | \n",
- " iteration | \n",
- " | \n",
- "
\n",
- " \n",
- " \n",
- " \n",
- " 0 | \n",
- " 0 | \n",
- " -23391.198596 | \n",
- "
\n",
- " \n",
- " 500 | \n",
- " -16429.246383 | \n",
- "
\n",
- " \n",
- " 1000 | \n",
- " -15210.110480 | \n",
- "
\n",
- " \n",
- " 1500 | \n",
- " -14895.824563 | \n",
- "
\n",
- " \n",
- " 2000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 2500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 3000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 3500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 4000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 4500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 5000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 5500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 6000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 6500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 7000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 7500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 8000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 8500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 9000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 9500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 10000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 10500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 11000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 11500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 12000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 12500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 13000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 13500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 14000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 14500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " ... | \n",
- " ... | \n",
- " ... | \n",
- "
\n",
- " \n",
- " 23 | \n",
- " 5000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 5500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 6000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 6500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 7000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 7500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 8000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 8500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 9000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 9500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 10000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 10500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 11000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 11500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 12000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 12500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 13000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 13500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 14000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 14500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 15000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 15500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 16000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 16500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 17000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 17500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 18000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 18500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 19000 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- " 19500 | \n",
- " -14681.308608 | \n",
- "
\n",
- " \n",
- "
\n",
- "
960 rows à 1 columns
\n",
- "
"
- ],
"text/plain": [
- " fitness\n",
- "worker iteration \n",
- "0 0 -23391.198596\n",
- " 500 -16429.246383\n",
- " 1000 -15210.110480\n",
- " 1500 -14895.824563\n",
- " 2000 -14681.308608\n",
- " 2500 -14681.308608\n",
- " 3000 -14681.308608\n",
- " 3500 -14681.308608\n",
- " 4000 -14681.308608\n",
- " 4500 -14681.308608\n",
- " 5000 -14681.308608\n",
- " 5500 -14681.308608\n",
- " 6000 -14681.308608\n",
- " 6500 -14681.308608\n",
- " 7000 -14681.308608\n",
- " 7500 -14681.308608\n",
- " 8000 -14681.308608\n",
- " 8500 -14681.308608\n",
- " 9000 -14681.308608\n",
- " 9500 -14681.308608\n",
- " 10000 -14681.308608\n",
- " 10500 -14681.308608\n",
- " 11000 -14681.308608\n",
- " 11500 -14681.308608\n",
- " 12000 -14681.308608\n",
- " 12500 -14681.308608\n",
- " 13000 -14681.308608\n",
- " 13500 -14681.308608\n",
- " 14000 -14681.308608\n",
- " 14500 -14681.308608\n",
- "... ...\n",
- "23 5000 -14681.308608\n",
- " 5500 -14681.308608\n",
- " 6000 -14681.308608\n",
- " 6500 -14681.308608\n",
- " 7000 -14681.308608\n",
- " 7500 -14681.308608\n",
- " 8000 -14681.308608\n",
- " 8500 -14681.308608\n",
- " 9000 -14681.308608\n",
- " 9500 -14681.308608\n",
- " 10000 -14681.308608\n",
- " 10500 -14681.308608\n",
- " 11000 -14681.308608\n",
- " 11500 -14681.308608\n",
- " 12000 -14681.308608\n",
- " 12500 -14681.308608\n",
- " 13000 -14681.308608\n",
- " 13500 -14681.308608\n",
- " 14000 -14681.308608\n",
- " 14500 -14681.308608\n",
- " 15000 -14681.308608\n",
- " 15500 -14681.308608\n",
- " 16000 -14681.308608\n",
- " 16500 -14681.308608\n",
- " 17000 -14681.308608\n",
- " 17500 -14681.308608\n",
- " 18000 -14681.308608\n",
- " 18500 -14681.308608\n",
- " 19000 -14681.308608\n",
- " 19500 -14681.308608\n",
- "\n",
- "[960 rows x 1 columns]"
+ "(-13000, -6600)"
]
},
- "execution_count": 6,
+ "execution_count": 17,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
- "trace = pd.read_csv('hillclimbing-random-trigram-uniform.csv').set_index(['worker', 'iteration']).sort_index()\n",
- "workers = list(sorted(set(trace.index.get_level_values(0))))\n",
- "trace"
+ "trigram_ylim = None\n",
+ "for f in glob.glob(\"*trigram*.csv\"):\n",
+ " df = pd.read_csv(f)\n",
+ " y1 = df.fitness.max()\n",
+ " y0 = df.fitness.min()\n",
+ " print(y0, y1, f)\n",
+ " if trigram_ylim:\n",
+ " trigram_ylim = (min(trigram_ylim[0], y0), max(trigram_ylim[1], y1))\n",
+ " else:\n",
+ " trigram_ylim = (y0, y1)\n",
+ "trigram_ylim = (math.floor(trigram_ylim[0] / 200) * 200, math.ceil(trigram_ylim[1] / 200) * 200)\n",
+ "trigram_ylim "
]
},
{
"cell_type": "code",
- "execution_count": 11,
+ "execution_count": null,
"metadata": {},
- "outputs": [
- {
- "data": {
- "image/png": "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\n",
- "text/plain": [
- "