League-of-Fortuneteller

기계학습을 이용해 게임(리그 오브 레전드) 결과를 예측해보자

Feature list

• blue_team_level
• blue_team_kill
• blue_team_death
• blue_team_assist
• blue_team_ckrate (킬 기여 수)
• blue_team_damage
• blue_team_bought_pink_ward
• blue_team_installed_ward
• blue_team_removed_ward
• blue_team_cs
• blue_team_cs_per_minute
• blue_team_gold
• blue_team_tier (Bronze, Silver 등 단계에 따라 숫자로 환산한 데이터)
• red_team_level
• red_team_kill
• red_team_death
• red_team_assist
• red_team_ckrate
• red_team_damage
• red_team_bought_pink_ward
• red_team_installed_ward
• red_team_removed_ward
• red_team_cs
• red_team_cs_per_minute
• red_team_gold
• red_team_tier

Model

• Dense(same size with input) with Activation ReLu
• Dense_1 with Activation Sigmoid

Result

Accuracy: 95.37%

   5/2289 [..............................] - ETA: 0s - loss: 0.0028 - acc: 1.0000
 150/2289 [>.............................] - ETA: 0s - loss: 0.0845 - acc: 0.9667
 305/2289 [==>...........................] - ETA: 0s - loss: 0.1004 - acc: 0.9639
 425/2289 [====>.........................] - ETA: 0s - loss: 0.0920 - acc: 0.9671
 570/2289 [======>.......................] - ETA: 0s - loss: 0.0971 - acc: 0.9614
 710/2289 [========>.....................] - ETA: 0s - loss: 0.1036 - acc: 0.9577
 845/2289 [==========>...................] - ETA: 0s - loss: 0.1042 - acc: 0.9598
1005/2289 [============>.................] - ETA: 0s - loss: 0.1038 - acc: 0.9612
1150/2289 [==============>...............] - ETA: 0s - loss: 0.1063 - acc: 0.9617
1285/2289 [===============>..............] - ETA: 0s - loss: 0.1059 - acc: 0.9603
1435/2289 [=================>............] - ETA: 0s - loss: 0.1044 - acc: 0.9610
1595/2289 [===================>..........] - ETA: 0s - loss: 0.1106 - acc: 0.9574
1755/2289 [======================>.......] - ETA: 0s - loss: 0.1127 - acc: 0.9561
1910/2289 [========================>.....] - ETA: 0s - loss: 0.1221 - acc: 0.9534
2075/2289 [==========================>...] - ETA: 0s - loss: 0.1193 - acc: 0.9533
2230/2289 [============================>.] - ETA: 0s - loss: 0.1192 - acc: 0.9534
2289/2289 [==============================] - 0s - loss: 0.1178 - acc: 0.9537   

Usage

Make training dataset

make_data.py is only for make training data.

FYI, our repo has already good train_data, so you can use it.

after make nickname_list.txt file with nickname list, run make_data.py.

(env3) $ python make_data.py

Then, if you see train_data folder, there will be two csv files. computer_trainable.csv and human_readable.csv.
you need computer_trainable.csv. so rename the computer_trainable.csv to data.csv.

Train model

model.py is our predict model. Just run model.py

(env3) $ python model.py
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
Epoch 1/20
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
...
...
...

Thanks