{"id":5826,"date":"2021-05-27T21:26:48","date_gmt":"2021-05-27T12:26:48","guid":{"rendered":"https:\/\/kats-eye.net\/info\/?p=5826"},"modified":"2021-05-29T08:22:26","modified_gmt":"2021-05-28T23:22:26","slug":"raspi-scikit-learn-svm","status":"publish","type":"post","link":"https:\/\/kats-eye.net\/info\/2021\/05\/27\/raspi-scikit-learn-svm\/","title":{"rendered":"\u30e9\u30ba\u30d1\u30a4\uff14 scikit-learn \u3067\u6a5f\u68b0\u5b66\u7fd2\uff08SVM\uff09\u3092\u5b66\u3076"},"content":{"rendered":"\n<h2>\u6982\u8981<\/h2>\n\n\n\n<p>\u3000\u524d\u56de\u307e\u3067\u306e\u6295\u7a3f\u3067\u306f\u3001\u691c\u51fa\u3057\u305f\u96fb\u8a71\u7740\u4fe1\u97f3\u306e\u5468\u6ce2\u6570\u6210\u5206\u3092\u4e00\u5b9a\u306e\u30eb\u30fc\u30eb\u306b\u57fa\u3065\u3044\u3066\u5224\u5b9a\u3057\u307e\u3057\u305f\u3002\u4eca\u56de\u306f\u3001\uff12\u3064\u306e\u96fb\u8a71\u7740\u4fe1\u97f3\u306e\u5206\u985e\u306bscikit-learn \u306e\u8a8d\u8b58\u30e2\u30c7\u30eb\u3067\u3042\u308bSVM\u3092\u5229\u7528\u3057\u307e\u3059\u3002<br><br>\u3000\u203b scikit-learn \uff1a \u6a5f\u68b0\u5b66\u7fd2\u30e9\u30a4\u30d6\u30e9\u30ea\u306e\u540d\u79f0<br>\u3000\u203b SVM \uff1a Support Vector Machine<\/p>\n\n\n\n<h2>SVM \uff08\u8abf\u67fb\u5185\u5bb9\uff09<\/h2>\n\n\n\n<p>\u3000\u6a5f\u68b0\u5b66\u7fd2\u306b\u304a\u3051\u308b\u5206\u985e\u3068\u56de\u5e30\u3092\u53d6\u308a\u6271\u3044\u3001\u6559\u5e2b\u3042\u308a\u5b66\u7fd2\u3092\u7528\u3044\u308b\u30d1\u30bf\u30fc\u30f3\u8a8d\u8b58\u30e2\u30c7\u30eb\u306e\u4e00\u3064\u3002\u57fa\u672c\u306f\u5b8c\u5168\u306b\u76f4\u7dda\u30fb\u5e73\u9762\u30fb\u8d85\u5e73\u9762\u3067\u7dda\u5f62\u5206\u96e2\u53ef\u80fd\u306a\u30c7\u30fc\u30bf\u3092\u30de\u30fc\u30b8\u30f3\u6700\u5927\u5316\u3068\u3044\u3046\u8003\u3048\u65b9\u306b\u57fa\u3065\u304d\uff12\u3064\u306e\u30af\u30e9\u30b9\u306b\u5206\u985e\u3059\u308b\u5883\u754c\u3092\u6c42\u3081\u308b\u3002\u4f46\u3057\u3001\u30bd\u30d5\u30c8\u30de\u30fc\u30b8\u30f3\u306b\u3088\u308b\u7dda\u5f62\u5206\u96e2\u51fa\u6765\u306a\u3044\u30c7\u30fc\u30bf\u306e\u5883\u754c\u8a2d\u5b9a\u3001\u66f2\u7dda\u30fb\u66f2\u9762\u5883\u754c\u8a2d\u5b9a\u3001\u8907\u6570\u30af\u30e9\u30b9\u3078\u306e\u5206\u985e\u7b49\u306b\u3082\u5bfe\u5fdc\u3059\u308b\u3068\u306e\u3053\u3068\u3002\u3000<br>\u3000\u5c1a\u3001\u6559\u5e2b\u3042\u308a\u5b66\u7fd2\u306f\u3001\u81a8\u5927\u306a\u5165\u529b\u30c7\u30fc\u30bf\uff08\u7279\u5fb4\u5024\uff09\u3068\u6b63\u89e3\u3092\u5b66\u7fd2\u3055\u305b\u308b\u3053\u3068\u3067\u3001\u30ed\u30b8\u30c3\u30af\u81ea\u8eab\u3092\u6539\u826f\u3059\u308b\u65b9\u6cd5\u3067\u3042\u308b\u3002<br><\/p>\n\n\n\n<h2>SCIKIT-LEARN \u30a4\u30f3\u30b9\u30c8\u30fc\u30eb<\/h2>\n\n\n\n<p>\u3000scikit-learn \u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3057\u307e\u3059\u3002<\/p>\n\n\n\n<div class=\"hcb_wrap\"><pre class=\"prism line-numbers lang-plain\"><code>sudo apt-get update\nsudo apt-get install python-sklearn<\/code><\/pre><\/div>\n\n\n\n<div class=\"wp-block-image\"><figure class=\"aligncenter size-large\"><img loading=\"lazy\" width=\"729\" height=\"715\" src=\"https:\/\/kats-eye.net\/info\/wp-content\/uploads\/2021\/05\/011.jpg\" alt=\"\" class=\"wp-image-5834\" srcset=\"https:\/\/kats-eye.net\/info\/wp-content\/uploads\/2021\/05\/011.jpg 729w, https:\/\/kats-eye.net\/info\/wp-content\/uploads\/2021\/05\/011-300x294.jpg 300w\" sizes=\"(max-width: 729px) 100vw, 729px\" \/><\/figure><\/div>\n\n\n\n<!-- START MoshimoAffiliateEasyLink --><script type=\"text\/javascript\">(function(b,c,f,g,a,d,e){b.MoshimoAffiliateObject=a;b[a]=b[a]||function(){arguments.currentScript=c.currentScript||c.scripts[c.scripts.length-2];(b[a].q=b[a].q||[]).push(arguments)};c.getElementById(a)||(d=c.createElement(f),d.src=g,d.id=a,e=c.getElementsByTagName(\"body\")[0],e.appendChild(d))})(window,document,\"script\",\"\/\/dn.msmstatic.com\/site\/cardlink\/bundle.js?20210203\",\"msmaflink\");msmaflink({\"n\":\"scikit-learn \u30c7\u30fc\u30bf\u5206\u6790 \u5b9f\u8df5\u30cf\u30f3\u30c9\u30d6\u30c3\u30af (Python\u30e9\u30a4\u30d6\u30e9\u30ea\u5b9a\u756a\u30bb\u30ec\u30af\u30b7\u30e7\u30f3)\",\"b\":\"\u79c0\u548c\u30b7\u30b9\u30c6\u30e0\",\"t\":\"\",\"d\":\"https:\\\/\\\/m.media-amazon.com\",\"c_p\":\"\",\"p\":[\"\\\/images\\\/I\\\/51BEBSEn4lL._SL500_.jpg\"],\"u\":{\"u\":\"https:\\\/\\\/www.amazon.co.jp\\\/dp\\\/4798055425\",\"t\":\"amazon\",\"r_v\":\"\"},\"v\":\"2.1\",\"b_l\":[{\"id\":1,\"u_tx\":\"Amazon\u3067\u898b\u308b\",\"u_bc\":\"#f79256\",\"u_url\":\"https:\\\/\\\/www.amazon.co.jp\\\/dp\\\/4798055425\",\"a_id\":1848374,\"p_id\":170,\"pl_id\":27060,\"pc_id\":185,\"s_n\":\"amazon\",\"u_so\":1},{\"id\":2,\"u_tx\":\"\u697d\u5929\u5e02\u5834\u3067\u898b\u308b\",\"u_bc\":\"#f76956\",\"u_url\":\"https:\\\/\\\/search.rakuten.co.jp\\\/search\\\/mall\\\/scikit-learn%20%E3%83%87%E3%83%BC%E3%82%BF%E5%88%86%E6%9E%90%20%E5%AE%9F%E8%B7%B5%E3%83%8F%E3%83%B3%E3%83%89%E3%83%96%E3%83%83%E3%82%AF%20(Python%E3%83%A9%E3%82%A4%E3%83%96%E3%83%A9%E3%83%AA%E5%AE%9A%E7%95%AA%E3%82%BB%E3%83%AC%E3%82%AF%E3%82%B7%E3%83%A7%E3%83%B3)\\\/\",\"a_id\":1848376,\"p_id\":54,\"pl_id\":27059,\"pc_id\":54,\"s_n\":\"rakuten\",\"u_so\":2},{\"id\":3,\"u_tx\":\"Yahoo!\u30b7\u30e7\u30c3\u30d4\u30f3\u30b0\u3067\u898b\u308b\",\"u_bc\":\"#66a7ff\",\"u_url\":\"https:\\\/\\\/shopping.yahoo.co.jp\\\/search?first=1\\u0026p=scikit-learn%20%E3%83%87%E3%83%BC%E3%82%BF%E5%88%86%E6%9E%90%20%E5%AE%9F%E8%B7%B5%E3%83%8F%E3%83%B3%E3%83%89%E3%83%96%E3%83%83%E3%82%AF%20(Python%E3%83%A9%E3%82%A4%E3%83%96%E3%83%A9%E3%83%AA%E5%AE%9A%E7%95%AA%E3%82%BB%E3%83%AC%E3%82%AF%E3%82%B7%E3%83%A7%E3%83%B3)\",\"a_id\":1848375,\"p_id\":1225,\"pl_id\":27061,\"pc_id\":1925,\"s_n\":\"yahoo\",\"u_so\":3}],\"eid\":\"cTIhF\",\"s\":\"xs\"});<\/script><div id=\"msmaflink-cTIhF\">\u30ea\u30f3\u30af<\/div><!-- MoshimoAffiliateEasyLink END -->\n\n\n\n<p>\u3000\u3000<br><\/p>\n\n\n\n<h2>\u30d7\u30ed\u30b0\u30e9\u30e0\u3010\u57fa\u790e\u3011<\/h2>\n\n\n\n<p>\u3000\u7c21\u5358\u306a\u30d7\u30ed\u30b0\u30e9\u30e0\u3067\u5b66\u7fd2\u30e2\u30c7\u30eb\u4f5c\u6210\u30fb\u4fdd\u5b58\u65b9\u6cd5\u3092\u78ba\u8a8d\u3057\u307e\u3059\u3002<\/p>\n\n\n\n<h4>\u2460 \u5b66\u7fd2\u30e2\u30c7\u30eb\u306e\u4f5c\u6210\u30fb\u4fdd\u5b58<\/h4>\n\n\n\n<p>\u3000\u884c\u756a2\uff653\u306f\u30e9\u30a4\u30d6\u30e9\u30ea\u3092\u30a4\u30f3\u30dd\u30fc\u30c8\u3057\u307e\u3059\u3002\u2018pickle\u2019 \u306f\u4f5c\u6210\u3057\u305f\u5b66\u7fd2\u30e2\u30c7\u30eb\u3092\u4fdd\u5b58\u3059\u308b\u70ba\u306b\u4f7f\u3044\u307e\u3059\u3002<br>\u3000\u884c\u756a6\u304c\u5165\u529b\u30c7\u30fc\u30bf\uff08\u7279\u5fb4\u5024\uff09\u3067\u3001\u884c\u756a7\u306f\u5165\u529b\u30c7\u30fc\u30bf\u306b\u5bfe\u5fdc\u3059\u308b\u6b63\u89e3\u3068\u306a\u308a\u307e\u3059\u3002\u305d\u308c\u305e\u308c\uff12\u3064\u306e\u7279\u5fb4\u5024\u3092\u6301\u3064\u30c7\u30fc\u30bf\u304b\u3089\uff12\u3064\u306e\u30af\u30e9\u30b9\uff08\u20180\u2019 or \u20181\u2019\uff09\u306b\u5206\u985e\u3057\u307e\u3059\u3002\u884c\u756a8\u3067 \u2018clf\u2019 \u3068\u3044\u3046\u5b66\u7fd2\u30e2\u30c7\u30eb\u306e\u5b66\u7fd2\u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3092\u8a2d\u5b9a\u3057\u3001\u884c\u756a9\u3067\u5b66\u7fd2\u30e2\u30c7\u30eb\u306b\u30c7\u30fc\u30bf\u30bb\u30c3\u30c8\u3092\u6e21\u3057\u3001\u5b66\u7fd2\u3055\u305b\u307e\u3059\u3002\u884c\u756a12\u306e\u691c\u8a3c\u30c7\u30fc\u30bf\u3092\u7528\u3044\u3066\u3001\u884c\u756a13\u3067\u5b66\u7fd2\u30e2\u30c7\u30eb \u2018clf\u2019 \u306b\u3088\u308b\u4e88\u6e2c\u7d50\u679c\u3092\u5f97\u307e\u3059\u3002\u3000\u884c\u756a16\uff6517\u3067\u4e88\u6e2c\u7d50\u679c\u3092\u51fa\u529b\u3057\u307e\u3059\u3002<br>\u3000\u884c\u756a21\uff6522\u306f\u3001\u518d\u5229\u7528\u76ee\u7684\u3067\u5b66\u7fd2\u30e2\u30c7\u30eb \u2018clf\u2019 \u3092\u4fdd\u5b58\u3057\u307e\u3059\u3002<\/p>\n\n\n\n<div class=\"hcb_wrap\"><pre class=\"prism line-numbers lang-python\" data-file=\"sk_learn_svm_s00_study_01.py\" data-lang=\"Python\"><code># -*- coding: utf-8 -*-\nfrom sklearn import svm\nimport pickle\n\n# \u5b66\u7fd2\ntrain_data = [[0, 0], [0, 1], [1, 0], [1, 1]]   # \u5165\u529b\u30c7\u30fc\u30bf\ntrain_label = [0, 0, 1, 1]                      # \u30e9\u30d9\u30eb\uff08\u6b63\u89e3\uff09\nclf = svm.SVC(C=10, gamma=0.1)                  # \u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3092\u6307\u5b9a\nclf.fit(train_data,train_label)                 # \u5b66\u7fd2\n\n# \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u306b\u5bfe\u3057\u3066\u4e88\u6e2c\ntest_data = [[0.6, 0], [1, 0], [0, 1], [0, 1]]  # \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\ntest_label = clf.predict(test_data)             # \u7d50\u679c\uff08\u4e88\u6e2c\u5024\uff09\n\n# \u30c6\u30b9\u30c8\u7d50\u679c\u306e\u8868\u793a\nprint(&quot;\u5165\u529b\u30c7\u30fc\u30bf\uff1a{0}&quot;.format(test_data))\nprint(&quot;\u4e88\u6e2c\u7d50\u679c  :{0}&quot;.format(test_label))\nprint(&quot;&quot;)\n\n# \u5b66\u7fd2\u30e2\u30c7\u30eb\u4fdd\u5b58\nwith open(&#39;model.pickle&#39;,mode=&#39;wb&#39;) as f :\n    pickle.dump(clf , f , protocol = 2 )<\/code><\/pre><\/div>\n\n\n\n<h4>\u2461 \u5b66\u7fd2\u30e2\u30c7\u30eb\u306e\u518d\u5229\u7528<\/h4>\n\n\n\n<p>\u3000\u2460\u3067\u4f5c\u6210\u3057\u305f\u5b66\u7fd2\u30e2\u30c7\u30eb\u3092\u518d\u5229\u7528\u3057\u307e\u3059\u3002\u884c\u756a2\u3067\u5b66\u7fd2\u30e2\u30c7\u30eb\u3092\u8aad\u307f\u8fbc\u3080\u969b\u306b\u5fc5\u8981\u306a \u2018pickle\u2019 \u3092\u30a4\u30f3\u30dd\u30fc\u30c8\u3057\u307e\u3059\u3002\u884c\u756a5\uff656 \u3067\u5b66\u7fd2\u306e\u6e08\u3093\u3060\u8a8d\u8b58\u30e2\u30c7\u30eb\u3092\u547c\u3073\u8fbc\u307f\u307e\u3059\u3002\u884c\u756a8 \u4ee5\u964d\u306f\u547c\u3073\u8fbc\u3093\u3060\u8a8d\u8b58\u30e2\u30c7\u30eb\u3092\u4f7f\u7528\u3057\u3001\u5165\u529b\u30c7\u30fc\u30bf\u306b\u5bfe\u3059\u308b\u4e88\u6e2c\u3092\u884c\u3063\u3066\u3044\u307e\u3059\u3002<\/p>\n\n\n\n<div class=\"hcb_wrap\"><pre class=\"prism line-numbers lang-python\" data-file=\"sk_learn_svm_s00_exe_01.py\" data-lang=\"Python\"><code># -*- coding: utf-8 -*-\nimport pickle\n\n# \u5b66\u7fd2\u30e2\u30c7\u30eb\u30ed\u30fc\u30c9\nwith open(&#39;model.pickle&#39;,mode=&#39;rb&#39;) as f :\n    clf = pickle.load(f)\n\n# \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u306b\u5bfe\u3057\u3066\u4e88\u6e2c\uff08\u305d\u306e\uff11\uff09\ntest_data = [[0.6, 0], [1, 0.9], [0, 1], [0, 1]]    # \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\ntest_label = clf.predict(test_data)                 # \u7d50\u679c\uff08\u4e88\u6e2c\u5024\uff09\n# \u4e88\u6e2c\u7d50\u679c\u8868\u793a\nprint(&quot;\u5165\u529b\u30c7\u30fc\u30bf\uff1a{0}&quot;.format(test_data))\nprint(&quot;\u4e88\u6e2c\u7d50\u679c  :{0}&quot;.format(test_label))\nprint(&quot;&quot;)\n\n# \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u306b\u5bfe\u3057\u3066\u4e88\u6e2c\uff08\u305d\u306e\uff12\uff09\ntest_data = [[0.2, 0], [0.5, 0.9]]                  # \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\ntest_label = clf.predict(test_data)                 # \u7d50\u679c\uff08\u4e88\u6e2c\u5024\uff09\n# \u4e88\u6e2c\u7d50\u679c\u8868\u793a\nprint(&quot;\u5165\u529b\u30c7\u30fc\u30bf\uff1a{0}&quot;.format(test_data))\nprint(&quot;\u4e88\u6e2c\u7d50\u679c  :{0}&quot;.format(test_label))\nprint(&quot;&quot;)\n\n# \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u306b\u5bfe\u3057\u3066\u4e88\u6e2c\uff08\u305d\u306e\uff13\uff09\ninp_1 = float(input(&quot;\u6570\u5b57\uff080--1\uff09\u5165\u529b\uff1a\u3000&quot;))\ninp_2 = float(input(&quot;\u6570\u5b57\uff080--1\uff09\u5165\u529b\uff1a\u3000&quot;))\n\ntemp_data = []\ntemp_data.append(inp_1)\ntemp_data.append(inp_2)\n\ntest_data = []\ntest_data.append(temp_data)                         # \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\n\ntest_label = clf.predict(test_data)                 # \u7d50\u679c\uff08\u4e88\u6e2c\u5024\uff09\nprint(&quot;\u5165\u529b\u30c7\u30fc\u30bf\uff1a{0}&quot;.format(test_data))\nprint(&quot;\u4e88\u6e2c\u7d50\u679c  :{0}&quot;.format(test_label))\nprint(&quot;&quot;)<\/code><\/pre><\/div>\n\n\n\n<p>\u3000\u3000<\/p>\n\n\n\n<h2>\u30d7\u30ed\u30b0\u30e9\u30e0\u3010\u767a\u5c55\u3011<\/h2>\n\n\n\n<h4>\u2460 \u7740\u4fe1\u97f3\u691c\u51fa\u7528 \u5b66\u7fd2\u30e2\u30c7\u30eb\u4f5c\u6210<\/h4>\n\n\n\n<p>\u3000\u524d\u56de\u6295\u7a3f\u3067\u3001\u5404\u7740\u4fe1\u97f3\u306e\u30d4\u30fc\u30af\u5468\u6ce2\u6570\u3092\u6b21\u306e\u69d8\u306bCSV\u5f62\u5f0f\u3067\u30d5\u30a1\u30a4\u30eb\u3057\u307e\u3057\u305f\u3002\u4eca\u56de\u3001\u3053\u306e\u7740\u4fe1\u97f3\uff3b01\uff3d\uff65\uff3b02\uff3d\u306eCSV\u30c7\u30fc\u30bf\u3092\u5165\u529b\u30c7\u30fc\u30bf\uff08\u7279\u5fb4\u5024\uff09\u3001\u5404\u7740\u4fe1\u97f3\u306b\u5bfe\u3059\u308b\u30e9\u30d9\u30eb\uff08\u6b63\u89e3\uff09\u3092\uff3b0\uff3d\uff3b1\uff3d\u3068\u3057\u3066\u8a8d\u8b58\u30e2\u30c7\u30eb\u306e\u5b66\u7fd2\u3092\u3057\u307e\u3059\u3002<\/p>\n\n\n\n<div class=\"hcb_wrap\"><pre class=\"prism line-numbers lang-plain\" data-file=\"SVM_01.csv\"><code>\u25c6\u7740\u4fe1\u97f3\uff3b01\uff3d\u30d4\u30fc\u30af\u5468\u6ce2\u6570\u3000CSV\u30d5\u30a1\u30a4\u30eb\u4fdd\u5b58\u4f8b\u25c6\n2654.639205460991,5303.6797234348\n2654.639205460991,5299.373040752351\n2653.777868924501,5300.234377288841\n2648.1791814373187,5302.81838689831\n2654.639205460991,5300.234377288841\n2654.639205460991,5304.110391703044\n\u3000\uff1c\u4ee5\u4e0b\u3001\u7701\u7565\uff08\u5168\uff12\uff10\uff11\u884c\uff09\uff1e<\/code><\/pre><\/div>\n\n\n\n<p>\u3000\u884c\u756a7\u301c30\u3067 \u7740\u4fe1\u97f3\uff3b01\uff3d\uff3b02\uff3d\u306e CSV\u30d5\u30a1\u30a4\u30eb\u3092\u7d9a\u3051\u3066\u8aad\u307f\u8fbc\u3093\u3067\u3001\u5165\u529b\u30c7\u30fc\u30bf\u3092\u914d\u5217 \u2018dt_frq_arr\u2019 \u306b\u3001\u6b63\u89e3\u3092 \u914d\u5217 \u2018dt_rtn_arr\u2019 \u306b\u683c\u7d0d\u3057\u307e\u3059\u3002<br>\u3000\u914d\u5217\u683c\u7d0d\u30c7\u30fc\u30bf\u3092\u5229\u7528\u3057\u3001\u884c\u756a33\uff6534\u3067\u5b66\u7fd2\u3057\u3001\u884c\u756a46\uff6547\u3067\u5b66\u7fd2\u7d50\u679c\u3092\u30d5\u30a1\u30a4\u30eb\u4fdd\u5b58\u3057\u307e\u3059\u3002<\/p>\n\n\n\n<div class=\"hcb_wrap\"><pre class=\"prism line-numbers lang-python\" data-file=\"sk_learn_svm_s01_study.py\" data-lang=\"Python\"><code># -*- coding: utf-8 -*-\nfrom sklearn import svm\nimport numpy as np\nimport pickle\nimport csv\n\ncsv_fil = [&#39;SVM_01.csv&#39;,&#39;SVM_02.csv&#39;]                               # CSV\u30d5\u30a1\u30a4\u30eb\u540d\ndt_frq_arr = np.empty((0,2) , int)                                  # \u5165\u529b\u30c7\u30fc\u30bf\u683c\u7d0d numpy.ndarray\ndt_rtn_arr = np.empty((0) , int)\u3000\u3000\u3000\u3000\u3000\u3000\u3000\u3000\u3000\u3000# \u30e9\u30d9\u30eb\uff08\u6b63\u89e3\uff09\u683c\u7d0d\u914d\u5217\n\nfor f_num in range(len(csv_fil)) :                                  # \u30d5\u30a1\u30a4\u30eb\u6570\u5206\u30eb\u30fc\u30d7\u51e6\u7406\n    dat_frg = []                                                    # \u4e00\u6642\u4fdd\u5b58\u30ea\u30b9\u30c8\n    with open(&#39;\/home\/pi\/MySndTest\/csv\/&#39; + csv_fil[f_num]) as f :    # CSV\u30d5\u30a1\u30a4\u30eb\u958b\u304f\n        reader = csv.reader(f)\n        for row in reader :                                         # \u30d5\u30a1\u30a4\u30eb\u884c\u30eb\u30fc\u30d7\n            if len(row)&gt;=2 :                                        # \uff11\u884c\u306b\uff12\u30c7\u30fc\u30bf\u4ee5\u4e0a\u3092\u5bfe\u8c61\n                tmp=[]\n                tmp.append(int(float(row[0])))                      # \u6587\u5b57\u5217\u3092int\u306b\u5909\u63db\n                tmp.append(int(float(row[1])))                      # \u6587\u5b57\u5217\u3092int\u306b\u5909\u63db  \n                dat_frg.append(tmp)                                 # \u4e00\u6642\u4fdd\u5b58\u30ea\u30b9\u30c8\u306b\u8ffd\u52a0\n\n                dt_rtn_arr = np.append(dt_rtn_arr,f_num)\n                \n\n    dt_frq = np.array(dat_frg)                                      # numpy.ndarray\u306b\u5909\u63db \n    dt_frq_ave = np.average(dt_frq , axis=0)                        # \u5e73\u5747\n    dt_frq_max = np.max(dt_frq , axis=0)                            # \u6700\u5927\n    dt_frq_min = np.min(dt_frq , axis=0)                            # \u6700\u5c0f\n\n    dt_frq_arr = np.append( dt_frq_arr , dt_frq , axis=0 )\n\n# \u5b66\u7fd2\nclf = svm.SVC(C=10 , gamma=0.1)                                 # \u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3092\u6307\u5b9a\nclf.fit(dt_frq_arr , dt_rtn_arr)                                # \u5b66\u7fd2\n\n# \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u306b\u5bfe\u3057\u3066\u4e88\u6e2c\ntest_data = [[2858,5684],[2841,5685],[2660,5296],[2642,5311]]   # \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\ntest_label = clf.predict(test_data)                             # \u7d50\u679c\uff08\u4e88\u6e2c\u5024\uff09\n\n# \u30c6\u30b9\u30c8\u7d50\u679c\u306e\u8868\u793a\nprint(&quot;\u5165\u529b\u30c7\u30fc\u30bf\uff1a{0}&quot;.format(test_data))\nprint(&quot;\u4e88\u6e2c\u7d50\u679c  :{0}&quot;.format(test_label))\nprint(&quot;&quot;)\n\n# \u5b66\u7fd2\u30e2\u30c7\u30eb\u4fdd\u5b58\npickle.dump(clf , open(&#39;frq_feature.pickle&#39;,mode=&#39;wb&#39;))\nprint(&quot;finished&quot;)<\/code><\/pre><\/div>\n\n\n\n<h4>\u2461 \u7740\u4fe1\u97f3\u691c\u51fa\u7528 \u5b66\u7fd2\u30e2\u30c7\u30eb\u8a66\u884c<\/h4>\n\n\n\n<p>\u3000\u4fdd\u5b58\u3057\u305f\u5b66\u7fd2\u30e2\u30c7\u30eb\u306e\u7c21\u6613\u30c6\u30b9\u30c8\u306b\u4f7f\u7528\u3057\u307e\u3057\u305f\u3002<\/p>\n\n\n\n<div class=\"hcb_wrap\"><pre class=\"prism line-numbers lang-python\" data-file=\"sk_learn_svm_s01_exec.py\" data-lang=\"Python\"><code># -*- coding: utf-8 -*-\nimport pickle\n\n# \u5b66\u7fd2\u30e2\u30c7\u30eb\u30ed\u30fc\u30c9\nclf = pickle.load(open(&#39;frq_feature.pickle&#39;,mode=&#39;rb&#39;))\n\n# \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u306b\u5bfe\u3057\u3066\u4e88\u6e2c\uff08\u305d\u306e\uff11\uff09\ntest_data = [[2858,5684],[2841,5685],[2660,5296],[2642,5311]]   # \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\ntest_label = clf.predict(test_data)                             # \u7d50\u679c\uff08\u4e88\u6e2c\u5024\uff09\n# \u4e88\u6e2c\u7d50\u679c\u8868\u793a\nprint(&quot;\u5165\u529b\u30c7\u30fc\u30bf\uff1a{0}&quot;.format(test_data))\nprint(&quot;\u4e88\u6e2c\u7d50\u679c  :{0}&quot;.format(test_label))\nprint(&quot;&quot;)\n\n# \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\u306b\u5bfe\u3057\u3066\u4e88\u6e2c\uff08\u305d\u306e\uff12\uff09\ninp_1 = float(input(&quot;\u5468\u6ce2\u6570\u5165\u529b\uff1a\u3000&quot;))\ninp_2 = float(input(&quot;\u5468\u6ce2\u6570\u5165\u529b\uff1a\u3000&quot;))\n\ntemp_data = []\ntemp_data.append(inp_1)\ntemp_data.append(inp_2)\n\ntest_data = []\ntest_data.append(temp_data)                         # \u30c6\u30b9\u30c8\u30c7\u30fc\u30bf\n\ntest_label = clf.predict(test_data)                 # \u7d50\u679c\uff08\u4e88\u6e2c\u5024\uff09\nprint(&quot;\u5165\u529b\u30c7\u30fc\u30bf\uff1a{0}&quot;.format(test_data))\nprint(&quot;\u4e88\u6e2c\u7d50\u679c  :{0}&quot;.format(test_label))\nprint(&quot;&quot;)<\/code><\/pre><\/div>\n\n\n\n<h4>\u2462 \u7740\u4fe1\u97f3\u691c\u51fa\u30d7\u30ed\u30b0\u30e9\u30e0<\/h4>\n\n\n\n<p>\u300011\u884c\u76ee\u3067\u4fdd\u5b58\u3057\u305f\u5b66\u7fd2\u30e2\u30c7\u30eb\u3092\u8aad\u307f\u8fbc\u307f\u307e\u3059\u3002\u884c\u756a71\u3067\u30c7\u30fc\u30bf\uff08\u7279\u5fb4\u5024\uff09\u3092\u5165\u529b\u3057\u3001\u7d50\u679c\uff08\u4e88\u6e2c\u5024\uff09\u3092\u53d6\u5f97\u3057\u307e\u3059\u3002\u96d1\u97f3\u306e\u5f71\u97ff\u3092\u53d7\u3051\u306b\u304f\u3044\u69d8\u306b\u3001\u884c\u756a57\u3067\u5404\u5468\u6ce2\u6570\u6210\u5206\u306e\u5927\u304d\u3055\u7b49\u3067\u30d5\u30a3\u30eb\u30bf\u30ea\u30f3\u30b0\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n\n\n\n<div class=\"hcb_wrap\"><pre class=\"prism line-numbers lang-python\" data-file=\"sk_learn_svm_s01_detect_sound.py\" data-lang=\"Python\"><code>import pyaudio\nimport wave\nimport numpy as np\nimport matplotlib.pyplot as plt\nimport csv\nimport os\nfrom scipy.signal import argrelmax\nimport pickle\n\n\nclf = pickle.load(open(&#39;frq_feature.pickle&#39;,mode=&#39;rb&#39;)) # \u5b66\u7fd2\u30e2\u30c7\u30eb\u30ed\u30fc\u30c9\n\nfrm_dt = pyaudio.paInt16                                # 16-bit resolution\nchnl = 1                                                # 1 channel\n\ns_rate = 44100                                          # 44.1kHz\u3000\u30b5\u30f3\u30d7\u30ea\u30f3\u30b0\u5468\u6ce2\u6570\nN = 100                                                 # chunk \/ s_rate = samp_time (sec)\nchunk = 1024 * N                                        # 2^12 \u4e00\u56de\u306e\u53d6\u5f97\u30c7\u30fc\u30bf\u6570 (1024*50\/44100= 1.16 (sec), 1024*100\/44100= 2.32 (sec))\ndev_idx = 5                                             # \u30c7\u30d0\u30a4\u30b9\u756a\u53f7\uff08\u5fc5\u9808\u3067\u306f\u306a\u3044\uff1f\uff08\u512a\u5148\u5ea6\u6307\u5b9a\u3057\u3066\u3044\u308b\u70ba\uff1f\uff09\uff09\n\naudio = pyaudio.PyAudio()                               # create pyaudio instantiation\n\n# \u30b9\u30c8\u30ea\u30fc\u30e0\u4f5c\u6210\nstream = audio.open(format = frm_dt , rate = s_rate , channels = chnl, input_device_index = dev_idx , \\\n                    input = True , frames_per_buffer = chunk )\n\n# x\u8ef8( frquency )\nwv_x1 = np.linspace(0, s_rate, chunk)\nchnk_lmt = int(chunk\/2)                                 # FFT\u6709\u52b9\u7bc4\u56f2\u8003\u616e\nwv_x2 = wv_x1[0:chnk_lmt]\n\n\n\n# \u97f3\u58f0\u53d6\u5f97\/FFT\u51e6\u7406\nprint(&quot;Started&quot;)\nwhile True:\n    try:\n        # \u97f3\u58f0\u30c7\u30fc\u30bf\u53d6\u5f97\n        data = stream.read(chunk , exception_on_overflow=False )\n        ndarray = np.frombuffer(data, dtype=&#39;int16&#39;)\n        \n        # y\u8ef8(FFT : amplitude)\n        wv_y1 = np.fft.fft(ndarray)\n        wv_y1 = np.abs(wv_y1)\n        wv_y2 = wv_y1[0:chnk_lmt]\n\n        peak_args = argrelmax(wv_y2,order=200)                  # \uff8b\uff9f\uff70\uff78\u691c\u51fa(argrelmax:\u6975\u5927\u5024\uff72\uff9d\uff83\uff9e\uff6f\uff78\uff7d ,order:\u6b21\u6975\u5927\u5024\u5224\u5b9a\u7bc4\u56f2\/\u9593\u9694(\u6700\u5c0f:int 1))\n\n        # peak_args(index) &gt;\u964d\u9806sort&gt; peak_args_sort(index)\u306b\u683c\u7d0d # \u25c6\u672a\u4f7f\u7528\uff08\u53c2\u8003\uff09\u25c6\n        f_peak = wv_y2[peak_args]                               # \uff72\uff9d\uff83\uff9e\uff6f\uff78\uff7d\u3092\u5f37\u5ea6\u5024\u306b\u5909\u63db\n        f_peak_argsort = f_peak.argsort()[::-1]                 # \u5f37\u5ea6\u964d\u9806\u4e26\u66ff(argsort\uff1a\u5024\u6607\u9806\uff7f\uff70\uff84(\uff72\uff9d\uff83\uff9e\uff6f\uff78\uff7d\u8fd4\u3059)\uff0c[::-1]\u964d\u9806\uff7f\uff70\uff84\u6642)\n        peak_args_sort = peak_args[0][f_peak_argsort]           # \u5f37\u5ea6\u964d\u9806\u4e26\u66ff\u5f8c\u306e\uff72\uff9d\uff83\uff9e\uff6f\uff78\uff7d\u914d\u5217\u3092\u6307\u5b9a\u3057\u3001\u5f37\u5ea6\u964d\u9806\u914d\u5217\u3092\u53d6\u5f97\n\n        # \u691c\u51fa\u30d4\u30fc\u30af\u306e\u5bfe\u8c61\u3092\u7d5e\u308a\u8fbc\u3080\n        wv_x3 = wv_x2[peak_args]                                 \n        wv_y3 = wv_y2[peak_args]\n        idx_tgt = np.where((wv_y3&gt;0.2*10**8) & (wv_x3&gt;=2500))   # \u6709\u52b9\u7bc4\u56f2\u8a2d\u5b9a\n\n        wv_x4 = wv_x3[idx_tgt]\n        wv_y4 = wv_y3[idx_tgt]\n\n        # \u30d4\u30fc\u30af\u30c7\u30fc\u30bf\u691c\u51fa\u6642\u3001\u5224\u5b9a\u51e6\u7406\n        msg = &quot;&quot;\n        if len(wv_x4) &gt; 1:\n            temp_data = []\n            temp_data.append(int(wv_x4[0]))\n            temp_data.append(int(wv_x4[1]))\n            \n            inp_data = []\n            inp_data.append(temp_data)                         # \u5165\u529b\u30c7\u30fc\u30bf\uff08\u691c\u51fa\u5468\u6ce2\u6570\uff09\n            rtn_label = clf.predict(inp_data)                  # \u7d50\u679c\uff08\u4e88\u6e2c\u5024\uff09\n            msg = &quot;[&quot;+(&quot;00&quot;+str(rtn_label[0]+1))[-2:]+&quot;]&quot; \n            \n        # \u30b0\u30e9\u30d5\u8868\u793a\u30c7\u30fc\u30bf\u8a2d\u5b9a\n        plt.plot(wv_x2 , wv_y2)                                 # FFT\u7d50\u679c \n        plt.plot(wv_x4 , wv_y4 ,&#39;ro&#39;)                           # \u30d4\u30fc\u30af\u5024\n\n        # \u30bf\u30a4\u30c8\u30eb\uff08\u5224\u5b9a\u7d50\u679c\u8868\u793a\uff09\n        if msg == &quot;&quot; :\n            plt.title(&quot;No phone is detected.&quot; , fontsize=16 , color=&#39;black&#39;)         \n        else :\n            plt.title(msg + &quot; is detected.&quot;, fontsize=22 , color=&#39;red&#39;) \n\n        plt.xlabel(&quot;frquency [Hz]&quot;)                             # X\u8ef8\u30e9\u30d9\u30eb\n        plt.ylabel(&quot;amplitude&quot;)                                 # Y\u8ef8\u30e9\u30d9\u30eb\n        plt.xlim(0, 10000)                                      # X\u8ef8\u8868\u793a\u7bc4\u56f2\n        plt.ylim(0, 3 * 10 ** 8)                                # Y\u8ef8\u8868\u793a\u7bc4\u56f2\n        \n        plt.minorticks_on()                                     # \u88dc\u52a9\u76ee\u76db\u308a\u8868\u793a\n        plt.grid(which=&quot;major&quot;, color=&quot;black&quot;, alpha=0.5)       # \u76ee\u76db\u308a\u7dda\u306e\u8868\u793a\n        plt.grid(which=&quot;minor&quot;, color=&quot;gray&quot;, linestyle=&quot;:&quot;)    # \u88dc\u52a9\u76ee\u76db\u308a\u7dda\u8868\u793a\n        \n        plt.draw()                                              # \u30ea\u30a2\u30eb\u30bf\u30a4\u30e0\u66f4\u65b0\n        plt.pause(0.0001)                                       # \u30ea\u30a2\u30eb\u30bf\u30a4\u30e0\u66f4\u65b0\n        plt.cla()                                               # \u73fe\u5728\u8ef8\u30af\u30ea\u30a2\n        \n    except KeyboardInterrupt:\n        print(&quot;Ctrl+C\u3067\u505c\u6b62\u3057\u307e\u3057\u305f&quot;)\n        break\n\nprint(&quot;Finished&quot;)\nplt.clf()\nplt.close()\n\n# \u30b9\u30c8\u30ea\u30fc\u30e0\u7d42\u4e86\nstream.stop_stream()\nstream.close()\naudio.terminate()\n<\/code><\/pre><\/div>\n\n\n\n<p>\u3000\u3000\u3000<\/p>\n\n\n\n<h2>\u691c\u51fa\u72b6\u6cc1\uff08\u52d5\u753b\uff09<\/h2>\n\n\n\n<p>\u3000\u30d7\u30ed\u30b0\u30e9\u30e0\u3010\u767a\u5c55\u3011\uff70 \u2462 \u7740\u4fe1\u97f3\u691c\u51fa\u30d7\u30ed\u30b0\u30e9\u30e0\u306e\u5b9f\u884c\u72b6\u6cc1\uff08\u52d5\u753b\uff09\u3067\u3059\u3002\u96d1\u97f3\u306e\u5f71\u97ff\u3092\u53d7\u3051\u306b\u304f\u3044\u5834\u6240\u3067\u306f\u3001\u5b89\u5b9a\u3057\u305f\u7d50\u679c\u3092\u5f97\u308b\u3053\u3068\u304c\u51fa\u6765\u307e\u3059\u3002\uff08\u52d5\u753b\u306b\u97f3\u58f0\u306f\u3064\u3051\u3066\u3044\u307e\u305b\u3093\u3002\uff09<\/p>\n\n\n\n<figure class=\"wp-block-video\"><video controls src=\"https:\/\/kats-eye.net\/info\/wp-content\/uploads\/2021\/05\/20210526b_SVM.mp4\"><\/video><\/figure>\n\n\n\n<p>\u3000\u3000\u3000<\/p>\n\n\n\n<h2>\u307e\u3068\u3081<\/h2>\n\n\n\n<p>\u3000\u5b9f\u7e3e\u30c7\u30fc\u30bf\u3092\u5165\u529b\u30fb\u5b66\u7fd2\u3055\u305b\u308b\u3068\u3001\u5224\u5b9a\u30ed\u30b8\u30c3\u30af\u3092\u81ea\u52d5\u3067\u4f5c\u6210\u3057\u3066\u304f\u308c\u308b\u306a\u3093\u3066\u3059\u3070\u3089\u3057\u3044\u3002<br>\u3000\u3067\u3082\u3001\u307e\u3063\u305f\u304f\u304b\u3051\u96e2\u308c\u305f\u30c7\u30fc\u30bf\u3092\u5165\u529b\u3057\u3066\u3082\u3069\u3061\u3089\u304b\u306b\u5224\u5b9a\u3059\u308b\u306e\u3060\u3068\u601d\u3044\u307e\u3059\u3002\u96d1\u97f3\u306e\u5f71\u97ff\u3092\u5c11\u306a\u304f\u3057\u305f\u308a\u3001\u3042\u308b\u7a0b\u5ea6\u5165\u529b\u524d\u306b\u30c7\u30fc\u30bf\u3092\u30c1\u30a7\u30c3\u30af\u3059\u308b\u306a\u3069\u306e\u691c\u8a0e\u3082\u5fc5\u8981\u3060\u3068\u601d\u3044\u307e\u3059\u3002<br>\u3000\u79c1\u3068\u3057\u3066\u306f\u6a5f\u68b0\u5b66\u7fd2\u306b\u3064\u3044\u3066\u306f\u521d\u3081\u3066\u306e\u3053\u3068\u306a\u306e\u3067\u3001\u3053\u308c\u304b\u3089\u3082\u3063\u3068\u826f\u3044\u65b9\u6cd5\u306b\u51fa\u4f1a\u3048\u308b\u3068\u601d\u3044\u307e\u3059\u3002<br><br><br>\u3000<\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u6982\u8981 \u3000\u524d\u56de\u307e\u3067\u306e\u6295\u7a3f\u3067\u306f\u3001\u691c\u51fa\u3057\u305f\u96fb\u8a71\u7740\u4fe1\u97f3\u306e\u5468\u6ce2\u6570\u6210\u5206\u3092\u4e00\u5b9a\u306e\u30eb\u30fc\u30eb\u306b\u57fa\u3065\u3044\u3066\u5224\u5b9a\u3057\u307e\u3057\u305f\u3002\u4eca\u56de\u306f\u3001\uff12\u3064\u306e\u96fb\u8a71\u7740\u4fe1\u97f3\u306e\u5206\u985e\u306bscikit-learn \u306e\u8a8d\u8b58\u30e2\u30c7\u30eb\u3067\u3042\u308bSVM\u3092\u5229\u7528\u3057\u307e\u3059\u3002 \u3000\u203b scikit-lea &hellip; <a href=\"https:\/\/kats-eye.net\/info\/2021\/05\/27\/raspi-scikit-learn-svm\/\" class=\"more-link\">\u7d9a\u304d\u3092\u8aad\u3080 <span class=\"screen-reader-text\">\u30e9\u30ba\u30d1\u30a4\uff14 scikit-learn \u3067\u6a5f\u68b0\u5b66\u7fd2\uff08SVM\uff09\u3092\u5b66\u3076<\/span> <span class=\"meta-nav\">&rarr;<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":5834,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[110],"tags":[164,116,165,166,167],"_links":{"self":[{"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/posts\/5826"}],"collection":[{"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/comments?post=5826"}],"version-history":[{"count":51,"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/posts\/5826\/revisions"}],"predecessor-version":[{"id":5905,"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/posts\/5826\/revisions\/5905"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/media\/5834"}],"wp:attachment":[{"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/media?parent=5826"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/categories?post=5826"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/kats-eye.net\/info\/wp-json\/wp\/v2\/tags?post=5826"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}