from google.colab import drive
drive.mount('/content/drive')

Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount("/content/drive", force_remount=True).

Load Data

import pandas as pd

train = pd.read_csv('/content/drive/MyDrive/Dacon/hand_gesture_data/train.csv')
test = pd.read_csv('/content/drive/MyDrive/Dacon/hand_gesture_data/test.csv')

train.drop(['id'], axis=1, inplace=True)
test.drop(['id'], axis=1, inplace=True)

EDA

train.head()

	sensor_1	sensor_2	sensor_3	sensor_4	sensor_5	sensor_6	sensor_7	sensor_8	sensor_9	sensor_10	...	sensor_24	sensor_25	sensor_26	sensor_27	sensor_28	sensor_29	sensor_30	sensor_31	sensor_32	target
0	-6.149463	-0.929714	9.058368	-7.017854	-2.958471	0.179233	-0.956591	-0.972401	5.956213	4.145636	...	-7.026436	-6.006282	-6.005836	7.043084	21.884650	-3.064152	-5.247552	-6.026107	-11.990822	1
1	-2.238836	-1.003511	5.098079	-10.880357	-0.804562	-2.992123	26.972724	-8.900861	-5.968298	-4.060134	...	-1.996714	-7.933806	-3.136773	8.774211	10.944759	9.858186	-0.969241	-3.935553	-15.892421	1
2	19.087934	-2.092514	0.946750	-21.831788	9.119235	17.853587	-21.069954	-15.933212	-9.016039	-5.975194	...	-6.889685	54.052330	-6.109238	12.154595	6.095989	-40.195088	-3.958124	-8.079537	-5.160090	0
3	-2.211629	-1.930904	21.888406	-3.067560	-0.240634	2.985056	-29.073369	0.200774	-1.043742	2.099845	...	-2.126170	-1.035526	2.178769	10.032723	-1.010897	-3.912848	-2.980338	-12.983597	-3.001077	1
4	3.953852	2.964892	-36.044802	0.899838	26.930210	11.004409	-21.962423	-11.950189	-20.933785	-4.000506	...	-2.051761	10.917567	1.905335	-13.004707	17.169552	2.105194	3.967986	11.861657	-27.088846	2

5 rows × 33 columns

train.corr()['target']

sensor_1    -0.031899
sensor_2    -0.029971
sensor_3    -0.059246
sensor_4    -0.014708
sensor_5     0.013416
sensor_6    -0.015927
sensor_7     0.023569
sensor_8    -0.004660
sensor_9    -0.005506
sensor_10    0.006381
sensor_11    0.013004
sensor_12    0.007906
sensor_13    0.019591
sensor_14    0.013936
sensor_15   -0.019540
sensor_16   -0.008993
sensor_17   -0.016523
sensor_18   -0.019564
sensor_19    0.001252
sensor_20   -0.019440
sensor_21    0.013673
sensor_22    0.004371
sensor_23    0.004909
sensor_24    0.003862
sensor_25    0.017889
sensor_26   -0.012798
sensor_27   -0.008405
sensor_28   -0.013656
sensor_29   -0.000169
sensor_30   -0.009233
sensor_31   -0.003516
sensor_32    0.041874
target       1.000000
Name: target, dtype: float64

df_data = train.drop(['target'], axis=1)
df_label = train.target

print(df_data.shape, df_label.shape)

(2335, 32) (2335,)

Modeling

from sklearn.model_selection import train_test_split
from sklearn.neighbors import KNeighborsClassifier
from sklearn.ensemble import RandomForestClassifier

X_train, X_test, y_train, y_test = train_test_split(df_data, df_label, test_size=0.25, random_state=1)

X_train.shape

(1751, 32)

• 데이터셋에 가장 잘 맞는 k값 찾기

import matplotlib.pyplot as plt

k_list = range(1, 101)
accuracy = list()

for k in k_list:
  knn = KNeighborsClassifier(n_neighbors=k)
  knn.fit(X_train, y_train)
  accuracy.append(knn.score(X_test, y_test))

plt.plot(k_list, accuracy)
plt.show()

k=3,

rf = RandomForestClassifier()
knn = KNeighborsClassifier(n_neighbors=3)

rf.fit(X_train, y_train)
knn.fit(X_train, y_train)

print(knn.score(X_test, y_test))
print(rf.score(X_test, y_test))

0.565068493150685
0.791095890410959

rf = RandomForestClassifier()
rf.fit(df_data, df_label)

RandomForestClassifier()

pre = rf.predict(test) 

Predict

sub = pd.read_csv('/content/drive/MyDrive/Dacon/hand_gesture_data/sample_submission.csv')
sub.target = pre
sub.to_csv('/content/drive/MyDrive/Dacon/hand_gesture_data/output1.csv', index=False)

sub.head()

	id	target
0	1	0
1	2	0
2	3	1
3	4	3
4	5	2