ãã¼ã¿åæã¬ãåå¼·ã¢ããã³ãã«ã¬ã³ãã¼ 24æ¥ç®ã
å½ãµã¤ãã§ããPythonã使ã£ããã¼ã¿åæãæ©æ¢°å¦ç¿ã«ã¤ãã¦ãåå¼·ããªãããããã¢ã¦ããããã¨ãã¦åºãã¨è¨ãããã¡ã§ãä½åãè¨äºãæ¸ãã¦ãã¾ããã è¨äºæ°ã§è¨ãã°50ã¨ããã®ããããããããªæ°ããã¾ãã
ã«ã¬ã³ãã¼ãå®æãã¤ã¤ããã®ã§ãåã ã§ã¯å½ãµã¤ãã®ç·ã¾ã¨ãã¨ãã¦ãæ©æ¢°å¦ç¿ããã¼ã¿åæã«è§¦ãããã¨ãã人ãã¼ãããå§ãã¦è§¦ããããããã«ãè¨äºãã¾ã¨ãã¦ããããã¨æãã¾ãã ä½ãé¢ç½ããã¨ãåå¼·ãããå¦çãå°±è·ã¾ã§ã®åå¼·ã«ãæ¥ã«æ©æ¢°å¦ç¿ã使ããªããã°ãªããªã社ä¼äººæ¹ã ã¯ãèªãã§è¦ã¦ãã ããã
- 0. ç°å¢æ§ç¯
- 1. Pythonã®ä½¿ãæ¹(åºæ¬ã©ã¤ãã©ãª)
- 2. ãã¼ã¿åæäºå§ã
- 3. ãã¼ã¿åæã®æµã
- 4. æç³»åãã¼ã¿
- 6. 深層å¦ç¿
- 7. æ師ãªãå¦ç¿(ç°å¸¸æ¤ç¥/次å åæ¸ãªã©)
- 8. å¼·åå¦ç¿
- 9. å®éã«ãã¼ã¿åæããã¦ããè¨äº
- ã¾ã¨ã
0. ç°å¢æ§ç¯
ã¾ãã¯ä½ã¨è¨ã£ã¦ãç°å¢æ§ç¯ãPythonã¯ä»ã®è¨èªã«æ¯ã¹ã¦æ§ç¯ãç°¡åãªæ¹ã ã¨ã¯æãã¾ãããããã§ããªãè¦æ¦ãã人ãå¤ãã§ãã
0.1. Pythonã®å°å ¥ (Anaconda)
å½ãµã¤ãã§ã¯Windowsã§ãç°å¢æ§ç¯ãç°¡åã«åºæ¥ãããã«è¨äºãæ¸ãã¦ãã¾ã *1*2ãå ¨ãã¼ãããå§ããæ¹ã¯ã³ãã©ãã©ããã
0.2. ã¨ãã£ã¿ (Pycharm/VSCode)
ããããé«æ©è½ã¡ã¢å¸³ãå®éã«ã³ã¼ããæ¸ãå§ããã¨ãèªå好ã¿ã®ã¨ãã£ã¿ã欲ãããªããã®ã§ããå½ãµã¤ãã§ã¯ãPythonãªãPycharmããã¾ããªã¼ã«ã©ã¦ã³ãã«ä½¿ããã¨ãã£ã¿ã¨ãã¦VSCodeãç´¹ä»ãã¦ãã¾ãã
0.3. ãã¼ã¸ã§ã³ç®¡ç (Git)
(â»ããã°ä¾¿å©ã§ããç¡ãã¦ãã³ã¼ãã¯æ¸ããã®ã§ãé¢åãªæ¹ã¯é£ã°ãã¦ããã£ã¦OKã§ãã)
è³¢ãã³ã¼ãã管çããä»çµã¿ã§ããã¡ãã£ã¨è©¦ããããã¨ãããã ããªã®ã«ãã³ã¼ããä¸é¨å¤æ´ãã¦ããã®ãééã£ã¦ä¿åãããããã¨æ²ããã§ãããã ãããªãã¨ããªãããã«ããã£ããã¨èªåã®ä½ã£ãã³ã¼ãã®ãã¼ã¸ã§ã³ç®¡çãã¾ããããGitã®åºæ¬çãªä½¿ãæ¹ã¯ãã¡ã
ã¾ããå½ãµã¤ãã®Githubã¢ã«ã¦ã³ãã¯ã³ãã©ã§ãã
ä½ã£ãããã¡ããåå¼·ã®æ§åãä¿ç®¡ãã¦ããã®ã§ããããããã°ãã²ã
1. Pythonã®ä½¿ãæ¹(åºæ¬ã©ã¤ãã©ãª)
ç°å¢ãä½ããããããããã³ã¼ããæ¸ãã¦ããã¾ãã
Pythonã¯ã©ã¤ãã©ãªãå
å®ãã¦ããã®ã§ããããã使ããªããã³ã¼ããæ¸ãã¦ããã¾ãã
æ©æ¢°å¦ç¿å°ç¨ã®ã©ã¤ãã©ãªãããã®ã§ãããã¾ãå¿
ãå
¥ãã¦ããã¹ãã©ã¤ãã©ãªã§ããnumpy
, matplotlib
, pandas
, jupyter
ã®ä½¿ãæ¹ãç´¹ä»ãã¦ãã¾ãã
1.1. æ°å¤è¨ç® : numpy
ãããç¨ãããã¨ã§é¢åãªåºæ¬è¨ç®(å¹³åè¨ç®ãè¡åè¨ç®ãetc...)ãä¸è¡ã§æ¸ããããã«ãªãã¾ããã¹ã´ãï¼
1.2. 表è¨ç®/çµ±è¨å¦ç : pandas
Pythonã§è¡¨ãèªã¿è¾¼ãã ããè¨ç®ããããã¾ã表ã¨ãã¦æ¸ãåºãããããã«ã¯pandasãã¨ã¦ã便å©ã å°ã使ãæ¹ã«ã¯ã»ãããã¾ãããããªãå¼·åãªã®ã§ãæ¯é身ã«ã¤ãã¦ãããããã®ã§ãã
1.3. ã°ã©ãæç» : matplotlib
ãã¼ã¿åæããã¦ããã¨ãçµæãã°ã©ãã§å¯è¦åãããã¨ãããã¨ãå¤ãããã¾ãã ããããã¨ãã«ã¯matplotlib. ããä¸ã¤ã§è²ã ãªã°ã©ããæ¸ããã¨ãåºæ¥ã¾ãã
ã¾ãããããªã·ã£ã¬ãªã°ã©ããä½ãããæ¹åãã«ãKibanaã¨Elasticsearchã§å¯è¦åããæ¹æ³ãæ¸ãã¦ãã¾ãããã®å ´åã¯ãPython以å¤ã®ç¥èãå¿ è¦ã§ãã
1.4. ã¤ã³ã¿ã©ã¯ãã£ããªã³ã¼ãå®è¡ jupyter
jupyterã®è¯ãç¹ã¯ã¡ãã£ã¨ãã¤ã³ã¼ããæ¸ããªãã試ããã¨ãåºæ¥ãã¨ããã¨ããã§ãã ã³ã¼ãæ¸ãã¦ãã³ã³ãã¤ã«ãã¦ãå®è¡...ãçããµã¤ã¯ã«ã§åãã¤ã¡ã¼ã¸ã åæã«ã¡ã¢ãæ®ããã®ã§ãç 究ãå®é¨ã«æé©ã§ãã
2. ãã¼ã¿åæäºå§ã
Pythonãããç¨åº¦ä½¿ããããã«ãªã£ããããã¼ã¿åæ/æ©æ¢°å¦ç¿ã®ãåå¼·ã§ãã
ãã®éã«åèã«ãªãæ¬ããµã¤ããã©ã¤ãã©ãªãªã©ããã£ã¨æ´ãåºããè¨äºãããã¾ãã®ã§ããªãã¡ã¬ã³ã¹ä»£ããã«ãã¦ããã ããã°ãªã¨æãã¾ãã
3. ãã¼ã¿åæã®æµã
ãããããå®éã«æ©æ¢°å¦ç¿ã交ããªããããã¼ã¿åæãé²ãã¾ãã
å®åã§ã¯ãä¸å³ã®ããã«ãã¼ã¿åæãé²ãã¦ããã®ã§ããããã®æµãããsklearnã使ããªãã説æãã¦ããã¾ãã
3.1 ãã¼ã¿ãéãã
ã¾ãããã¼ã¿ãéããªãã¨ããã¾ããã ã¢ãã«/ã¢ã«ã´ãªãºã ã®æ§è½ã調ã¹ããã¨ãã«ãsklearnã¯ãããããªãã¼ã¿ãäºãç¨æãã¦ããã¦ãã¾ãã
ã¾ããæ
å ±ãéãããã¨ãã¯ã¹ã¯ã¬ã¤ãã³ã°/ã¯ãã¼ãªã³ã°ãªã©ã®æè¡ãå¿
è¦ã§ãã
beautifulsoup
ã使ã£ãããå
¬éAPIã使ããªã©ãã¦ããã¼ã¿ãéãã¾ãããã
3.2 ãã¼ã¿åæã®æµã
4. æç³»åãã¼ã¿
æéä¾åã®ãããã¼ã¿ã¯ããã§ãªããã¼ã¿ã«å¯¾ãã¦ãå°ãæ±ããé¢åã ã£ãããã¾ãã ããã©ãæ±ããã¨ãå¤ãã®ã§ãã¾ã¨ãã¦ãã¾ããæç³»åã«é¢ããåºç¤çãªã¾ã¨ãããããã£ã¼ãã©ã¼ãã³ã°ã«è³ãã¾ã§ãã¾ã¨ã¾ã£ã¦ã¾ãã®ã§ãèªãã§ã¿ã¦ãã ããã
6. 深層å¦ç¿
æè¿çéãè³ããã¦ãããã£ã¼ãã©ã¼ãã³ã°ã«è§¦ããªã訳ã«ã¯ãããªãã§ãããã å½ãµã¤ãã§ã¯ãKerasã¨PyTorchã®2ã¤ã«çµã£ã¦ããã®ä½¿ãæ¹ã«ã¤ãã¦èª¬æãã¦ãã¾ãã
6.1 èªä½
ã©ã¤ãã©ãªã使ãåã«ããã£ã¼ãã©ã¼ãã³ã°ã®ã³ã³ããç¥ããããã¾ãã¯Pythonã§èªä½ãã¾ããã
1ããæ¸ãã¨ãã©ãããææ³ã§ãããã¯ã¼ã¯ãçµã¾ãã¦ããã®ãã¨ããå¦ç¿ã®ä»çµã¿ãªã©ãåããããã«ãªãã¾ãã
6.2. Keras
ã¨ã£ã¤ããããã§ã¯éä¸ã®Kerasãç°¡åãªãã¼ã¿åæã«ä½¿ãã«ã¯ããã§ååã ã¨æãã¾ãããã£ã¼ãã©ã¼ãã³ã°ãªãã¦ç¥ããªãã¦ã使ãããã®ãæ軽ãã¯ã¹ã´ãã
6.3 PyTorch
ç 究è çéã§çãä¸ãããè¦ãã¦ããPyTorchããããã³ã¼ãè¨è¿°éã¯å¢ãã¾ãããDefine by Runã®è¨è¨ããããªã®ã§ãããªãæè»ãªè¨è¨ãå¯è½ã æ´ã«ãSklearnã¨ã®é£æºã§ããã使ãããããªãã¾ããåã©ã¤ãã©ãªã¨ã®æ¯è¼ãæ¸ãã¦ãã¾ãã®ã§ãæ¯éãä¸èªãã
7. æ師ãªãå¦ç¿(ç°å¸¸æ¤ç¥/次å åæ¸ãªã©)
ä¸è¨ã¾ã§ããæ£è§£ãã¼ã¿ã®ã©ãã«ã«å¾ã£ã¦å¦ç¿ãé²ãã¦ããæ師ããå¦ç¿ã§ãã å®éã®ç¾å ´ã§ã¯æ£è§£ãã¼ã¿ã使ããªãã¿ã¹ã¯ãå¤ãããã¾ãã
ç°å¸¸æ¤ç¥*3ãã座æ¨ãå¤æãã空éå¤æãªã©ãæ£è§£ãã¼ã¿ã¨ããæ¦å¿µãç¡ããããªãã®ã«å¯¾ãã¦ãã¾ã¨ãã¾ããã®ã§ãå¿ è¦ãªæ¹ã¯ã©ããã
8. å¼·åå¦ç¿
å¼·åå¦ç¿ã¯ãæ師ããå¦ç¿ã¨ãæ師ãªãå¦ç¿ã¨ãéãã¾ãã
å®é¨ãã¦ããã®çµæã帰ã£ã¦ãã¦ãå¦ç¿ãã¦ã¾ãå®é¨ããã
ãã®ãããªã¾ãã§äººéã®ãããªå¦ç¿ã®æ¹æ³ã§ããå®å¼åãåºæ¥ãªãè¤éãªç¶æ³ãªã©ã«å¿ç¨ãæå¾ ããã¦ããã å®éã«ãã¼ã±ãã£ã³ã°ã®åºåã®é ç½®ãªã©ã«ä½¿ãããããã¦ãã¾ãã
9. å®éã«ãã¼ã¿åæããã¦ããè¨äº
å½ãµã¤ãã§ã¯ãå®éã«åæãè¡ã£ãè¨äºãããã¤ãè¼ãã¦ãã¾ãã
ããã¾ã§é«åº¦ãªãã¨ã¯ãã¦ãã¾ããããè¨äºãããã¼ã¿åæããåãããã¨ããã¼ã¿åæã®ã¡ãããæãã¦ããããã°ã¨æãã¾ãã
ã¾ã¨ã
ãããã§ããããããªããããããªãã¨ããã£ã¦ããã®ã§ãä¸ã¤ã«ã¾ã¨ãã¾ããã
ä»åã®ã¢ããã³ãã«ã¬ã³ãã¼ã¯ããã®è¨äºã®é çªãæèãã¦é²ãã¦ãã¾ããããã®æµãã§åå¼·ãã¦ããã¨ãå²ã¨ã¹ã ã¼ãºã«é²ããããã®ã§ã¯ãªãã§ããããã
ãã¼ã¿åææ¥çãçãä¸ãã£ã¦ããã®ã§ããããæ©ã«ä¸å¿µçºèµ·ãã¦åå¼·ãããæ¹ãªã©ãããã²ãåèãã ãããã§ã¯ã§ã¯ï¼
*1:MacãLinuxã¯ç°å¢æ§ç¯æ¹æ³ãWebã§æ¼ãã°ç°¡åã«åºã¦æ¥ãããã¾ããããã©ããããããªã«å¤ããªãã®ã§ãããã£ã¦ã¿ã¦ãã ããã
*2:æ©æ¢°å¦ç¿ã©ã¤ãã©ãªã«ãã£ã¦ã¯ãã©ããã¦ãwindowsã§åºæ¥ãªããã®ãããã¾ãã
*3:å³å¯ã«è¨ãã°æ師ã®ããç°å¸¸æ¤ç¥ãããã¾ãã