本文整理汇总了Python中sklearn.cluster.MeanShift.score_samples方法的典型用法代码示例。如果您正苦于以下问题：Python MeanShift.score_samples方法的具体用法？Python MeanShift.score_samples怎么用？Python MeanShift.score_samples使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类sklearn.cluster.MeanShift的用法示例。
在下文中一共展示了MeanShift.score_samples方法的1个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: analysis
# 需要导入模块: from sklearn.cluster import MeanShift [as 别名]
# 或者: from sklearn.cluster.MeanShift import score_samples [as 别名]
def analysis(lon, lat, day=None, hour=None, distance=1, method=1):
global LOAD_TAXI_DF, taxi_df
if not LOAD_TAXI_DF:
filename = file_path('nyc_data.csv')
taxi_df = pd.read_csv(filename, parse_dates=['pickup_datetime', 'dropoff_datetime'])
LOAD_TAXI_DF = True
columns = ['medallion', 'pickup_datetime', 'dropoff_datetime',
'pickup_latitude', 'pickup_longitude',
'dropoff_latitude', 'dropoff_longitude']
taxi_df = taxi_df[columns]
target_lat, target_lon = float(lat) , float(lon)
target_day = datetime.now().weekday() if not day else int(day)
target_hour = datetime.now().hour if not hour else int(hour)
width = 0.010 * float(distance)
lat_max, lon_max = target_lat + width, target_lon - width
lat_min, lon_min = target_lat - width, target_lon + width
hour_day_df = taxi_df[ (taxi_df.pickup_datetime.dt.hour==target_hour) & (taxi_df.pickup_datetime.dt.day==target_day)][['pickup_latitude', 'pickup_longitude']]
hour_day_df = hour_day_df[(hour_day_df['pickup_latitude'] > lat_min) & (hour_day_df['pickup_latitude'] < lat_max)]
hour_day_df = hour_day_df[(hour_day_df['pickup_longitude'] < lon_min) & (hour_day_df['pickup_longitude'] > lon_max)]
if method == 'MeanShift':
clf = MeanShift().fit(hour_day_df.values)
centers = clf.cluster_centers_
elif method == 'KernelDensity':
clf = KernelDensity(kernel='gaussian', bandwidth=0.005).fit(hour_day_df.values)
score = clf.score_samples(hour_day_df.values)
index_score = np.argsort(score)
centers = hour_day_df[index_score < 20].values
else:
clf = AffinityPropagation().fit(hour_day_df.values)
centers = clf.cluster_centers_
markers = [{'lon':str(center[0]), 'lat':str(center[1]) } for center in centers]
return markers