技能 aeon
T
時間序列數據需要專門的演算法，而非標準的機器學習方法。Aeon 提供與 scikit-learn 相容的 API，可用於對時間序列數據進行分類、回歸、聚類、預測、異常偵測、分割和相似性搜尋。
支援: Claude Codex Code(CC)
1
下載技能 ZIP
2
在 Claude 中上傳
前往 設定 → 功能 → 技能 → 上傳技能
3
開啟並開始使用
測試它
正在使用「aeon」。 使用 aeon 對我的時間序列資料進行分類
預期結果:
- RocketClassifier 訓練成功
- 測試集準確率：92.3%
- 關鍵參數：n_kernels=10000
- 資料形狀驗證：(samples, channels, timepoints)
安全審計
安全v4 • 1/17/2026
All 531 static findings are false positives. This is a documentation-only skill containing SKILL.md and references/*.md files with no executable code. The scanner incorrectly flagged markdown syntax (backticks for inline code), Python import examples in documentation, ML algorithm names (DTW, LCSS, ERP) misinterpreted as cryptographic references, and legitimate documentation URLs. No actual code execution, network calls, or credential access exists.
14
已掃描檔案
5,466
分析行數
4
發現項
4
審計總數
風險因素
⚡ 包含腳本 (8)
evaluation_output.json:33 references/anomaly_detection.md:125-130 references/datasets_benchmarking.md:176-181 references/datasets_benchmarking.md:200-203 references/datasets_benchmarking.md:217-220 references/forecasting.md:111-115 references/segmentation.md:131-134 references/transformations.md:168-172
🌐 網路存取 (6)
⚙️ 外部命令 (473)
references/anomaly_detection.md:9 references/anomaly_detection.md:13 references/anomaly_detection.md:25 references/anomaly_detection.md:29 references/anomaly_detection.md:33 references/anomaly_detection.md:37 references/anomaly_detection.md:41 references/anomaly_detection.md:45 references/anomaly_detection.md:49 references/anomaly_detection.md:57 references/anomaly_detection.md:61 references/anomaly_detection.md:69 references/anomaly_detection.md:73 references/anomaly_detection.md:77 references/anomaly_detection.md:83 references/anomaly_detection.md:89-107 references/anomaly_detection.md:107-124 references/anomaly_detection.md:124-136 references/classification.md:9 references/classification.md:10 references/classification.md:11 references/classification.md:12 references/classification.md:13 references/classification.md:21 references/classification.md:22 references/classification.md:23 references/classification.md:24 references/classification.md:25 references/classification.md:26 references/classification.md:27 references/classification.md:35 references/classification.md:36 references/classification.md:37 references/classification.md:38 references/classification.md:46 references/classification.md:47 references/classification.md:48 references/classification.md:56 references/classification.md:57 references/classification.md:58 references/classification.md:59 references/classification.md:60 references/classification.md:68 references/classification.md:69 references/classification.md:70 references/classification.md:71 references/classification.md:72 references/classification.md:73 references/classification.md:81 references/classification.md:82 references/classification.md:83 references/classification.md:84 references/classification.md:92 references/classification.md:93 references/classification.md:101 references/classification.md:102 references/classification.md:110 references/classification.md:118 references/classification.md:119 references/classification.md:120 references/classification.md:124-136 references/clustering.md:9 references/clustering.md:10 references/clustering.md:11 references/clustering.md:12 references/clustering.md:20 references/clustering.md:21 references/clustering.md:29 references/clustering.md:30 references/clustering.md:38 references/clustering.md:46 references/clustering.md:47 references/clustering.md:48 references/clustering.md:49 references/clustering.md:50 references/clustering.md:51 references/clustering.md:59 references/clustering.md:60 references/clustering.md:61 references/clustering.md:69 references/clustering.md:75 references/clustering.md:76 references/clustering.md:77 references/clustering.md:78 references/clustering.md:84-99 references/datasets_benchmarking.md:10-19 references/datasets_benchmarking.md:19-22 references/datasets_benchmarking.md:22-31 references/datasets_benchmarking.md:31-34 references/datasets_benchmarking.md:34-39 references/datasets_benchmarking.md:39-42 references/datasets_benchmarking.md:42-46 references/datasets_benchmarking.md:46-51 references/datasets_benchmarking.md:51-55 references/datasets_benchmarking.md:55-58 references/datasets_benchmarking.md:58-62 references/datasets_benchmarking.md:62-65 references/datasets_benchmarking.md:65-69 references/datasets_benchmarking.md:69-72 references/datasets_benchmarking.md:72-76 references/datasets_benchmarking.md:76-79 references/datasets_benchmarking.md:79-83 references/datasets_benchmarking.md:83-88 references/datasets_benchmarking.md:88-92 references/datasets_benchmarking.md:92-95 references/datasets_benchmarking.md:95-99 references/datasets_benchmarking.md:99-106 references/datasets_benchmarking.md:106-107 references/datasets_benchmarking.md:107-108 references/datasets_benchmarking.md:108-109 references/datasets_benchmarking.md:109-113 references/datasets_benchmarking.md:113-122 references/datasets_benchmarking.md:122-128 references/datasets_benchmarking.md:128-134 references/datasets_benchmarking.md:134-142 references/datasets_benchmarking.md:142-153 references/datasets_benchmarking.md:153-159 references/datasets_benchmarking.md:159-168 references/datasets_benchmarking.md:168-175 references/datasets_benchmarking.md:175-188 references/datasets_benchmarking.md:188-191 references/datasets_benchmarking.md:191-196 references/datasets_benchmarking.md:196-199 references/datasets_benchmarking.md:199-210 references/datasets_benchmarking.md:210-216 references/datasets_benchmarking.md:216-227 references/datasets_benchmarking.md:227-235 references/datasets_benchmarking.md:235-240 references/datasets_benchmarking.md:240-244 references/datasets_benchmarking.md:244-247 references/datasets_benchmarking.md:247-261 references/datasets_benchmarking.md:261-285 references/datasets_benchmarking.md:285-293 references/datasets_benchmarking.md:293-301 references/datasets_benchmarking.md:301-307 references/datasets_benchmarking.md:307-310 references/datasets_benchmarking.md:310-316 references/datasets_benchmarking.md:316-322 references/datasets_benchmarking.md:322-328 references/datasets_benchmarking.md:328-337 references/datasets_benchmarking.md:337-343 references/datasets_benchmarking.md:343-353 references/datasets_benchmarking.md:353-359 references/datasets_benchmarking.md:359-369 references/datasets_benchmarking.md:369-375 references/datasets_benchmarking.md:375-387 references/distances.md:12 references/distances.md:13 references/distances.md:14 references/distances.md:15 references/distances.md:16 references/distances.md:19 references/distances.md:20 references/distances.md:21 references/distances.md:22 references/distances.md:25 references/distances.md:26 references/distances.md:27 references/distances.md:35 references/distances.md:36 references/distances.md:37 references/distances.md:38 references/distances.md:46-54 references/distances.md:54-58 references/distances.md:58-67 references/distances.md:67-71 references/distances.md:71-80 references/distances.md:80-84 references/distances.md:84-94 references/distances.md:94-102 references/distances.md:102-108 references/distances.md:108-114 references/distances.md:114-117 references/distances.md:117-121 references/distances.md:121-130 references/distances.md:130-169 references/distances.md:169-174 references/distances.md:174-180 references/distances.md:180-183 references/distances.md:183-194 references/distances.md:194-197 references/distances.md:197-208 references/distances.md:208-220 references/distances.md:220-226 references/distances.md:226-236 references/distances.md:236-242 references/distances.md:242-247 references/distances.md:247-251 references/distances.md:251-256 references/forecasting.md:9 references/forecasting.md:10 references/forecasting.md:10 references/forecasting.md:18 references/forecasting.md:19 references/forecasting.md:19 references/forecasting.md:19 references/forecasting.md:23 references/forecasting.md:24 references/forecasting.md:24 references/forecasting.md:24 references/forecasting.md:28 references/forecasting.md:29 references/forecasting.md:33 references/forecasting.md:34 references/forecasting.md:34 references/forecasting.md:38 references/forecasting.md:45 references/forecasting.md:49 references/forecasting.md:57 references/forecasting.md:58 references/forecasting.md:58 references/forecasting.md:63-80 references/forecasting.md:80-84 references/forecasting.md:84-86 references/forecasting.md:86-93 references/forecasting.md:93-110 references/forecasting.md:110-121 references/forecasting.md:121-127 references/forecasting.md:127-133 references/forecasting.md:133-137 references/forecasting.md:137-138 references/networks.md:107-123 references/networks.md:123-129 references/networks.md:129-138 references/networks.md:138-142 references/networks.md:142-143 references/networks.md:143-144 references/networks.md:144-145 references/networks.md:145-155 references/networks.md:155-156 references/networks.md:156-157 references/networks.md:157-164 references/networks.md:164-174 references/networks.md:174-180 references/networks.md:180-187 references/networks.md:187-219 references/networks.md:219-225 references/networks.md:225-231 references/networks.md:231-240 references/networks.md:240-254 references/networks.md:254-266 references/networks.md:266-280 references/networks.md:280-289 references/regression.md:9 references/regression.md:10 references/regression.md:11 references/regression.md:12 references/regression.md:13 references/regression.md:21 references/regression.md:22 references/regression.md:23 references/regression.md:24 references/regression.md:25 references/regression.md:26 references/regression.md:27 references/regression.md:28 references/regression.md:29 references/regression.md:37 references/regression.md:45 references/regression.md:46 references/regression.md:47 references/regression.md:48 references/regression.md:56 references/regression.md:64 references/regression.md:65 references/regression.md:66 references/regression.md:67 references/regression.md:68 references/regression.md:69 references/regression.md:77 references/regression.md:85 references/regression.md:86 references/regression.md:87 references/regression.md:91 references/regression.md:92 references/regression.md:93 references/regression.md:97-109 references/segmentation.md:8 references/segmentation.md:10 references/segmentation.md:10 references/segmentation.md:14 references/segmentation.md:20 references/segmentation.md:26 references/segmentation.md:31 references/segmentation.md:37 references/segmentation.md:42 references/segmentation.md:47 references/segmentation.md:52 references/segmentation.md:57-73 references/segmentation.md:73-79 references/segmentation.md:79-82 references/segmentation.md:82-99 references/segmentation.md:99-104 references/segmentation.md:104-109 references/segmentation.md:109-114 references/segmentation.md:114-119 references/segmentation.md:119-124 references/segmentation.md:124-130 references/segmentation.md:130-141 references/segmentation.md:141-154 references/segmentation.md:154-163 references/similarity_search.md:10 references/similarity_search.md:16 references/similarity_search.md:22 references/similarity_search.md:31 references/similarity_search.md:38-58 references/similarity_search.md:58-62 references/similarity_search.md:62-79 references/similarity_search.md:79-83 references/similarity_search.md:83-97 references/similarity_search.md:97-108 references/similarity_search.md:108-121 references/similarity_search.md:121-136 references/similarity_search.md:136-141 references/similarity_search.md:141-146 references/similarity_search.md:146-150 references/similarity_search.md:150-155 references/similarity_search.md:155-162 references/transformations.md:17 references/transformations.md:18 references/transformations.md:19 references/transformations.md:20 references/transformations.md:21 references/transformations.md:22 references/transformations.md:30 references/transformations.md:31 references/transformations.md:32 references/transformations.md:33 references/transformations.md:41 references/transformations.md:42 references/transformations.md:43 references/transformations.md:44 references/transformations.md:45 references/transformations.md:46 references/transformations.md:54 references/transformations.md:55 references/transformations.md:56 references/transformations.md:57 references/transformations.md:65 references/transformations.md:66 references/transformations.md:67 references/transformations.md:75 references/transformations.md:76 references/transformations.md:77 references/transformations.md:78 references/transformations.md:79 references/transformations.md:80 references/transformations.md:88 references/transformations.md:89 references/transformations.md:90 references/transformations.md:91 references/transformations.md:92 references/transformations.md:93 references/transformations.md:97 references/transformations.md:98 references/transformations.md:99 references/transformations.md:105 references/transformations.md:113 references/transformations.md:114 references/transformations.md:115 references/transformations.md:119 references/transformations.md:120 references/transformations.md:121 references/transformations.md:122 references/transformations.md:123 references/transformations.md:124 references/transformations.md:130 references/transformations.md:131 references/transformations.md:135 references/transformations.md:136 references/transformations.md:137 references/transformations.md:141 references/transformations.md:145-163 references/transformations.md:163-167 references/transformations.md:167-181 references/transformations.md:181-185 references/transformations.md:185-191 references/transformations.md:191-216 references/transformations.md:216-220 references/transformations.md:220-223 references/transformations.md:223-229 references/transformations.md:229-232 references/transformations.md:232-236 SKILL.md:28-30 SKILL.md:30-36 SKILL.md:36-39 SKILL.md:39-51 SKILL.md:51-54 SKILL.md:54 SKILL.md:54-55 SKILL.md:55 SKILL.md:55-56 SKILL.md:56 SKILL.md:56-57 SKILL.md:57-61 SKILL.md:61-64 SKILL.md:64-74 SKILL.md:74-78 SKILL.md:78-81 SKILL.md:81-91 SKILL.md:91-95 SKILL.md:95-98 SKILL.md:98-104 SKILL.md:104-108 SKILL.md:108-111 SKILL.md:111-120 SKILL.md:120-124 SKILL.md:124-127 SKILL.md:127-132 SKILL.md:132-136 SKILL.md:136-139 SKILL.md:139-145 SKILL.md:145-149 SKILL.md:149-152 SKILL.md:152-162 SKILL.md:162-165 SKILL.md:165-170 SKILL.md:170-173 SKILL.md:173-178 SKILL.md:178-182 SKILL.md:182-185 SKILL.md:185-202 SKILL.md:202-211 SKILL.md:211-214 SKILL.md:214 SKILL.md:214 SKILL.md:214-215 SKILL.md:215 SKILL.md:215-216 SKILL.md:216 SKILL.md:216-219 SKILL.md:219-225 SKILL.md:225-229 SKILL.md:229-232 SKILL.md:232-240 SKILL.md:240-243 SKILL.md:243-248 SKILL.md:248-254 SKILL.md:254-266 SKILL.md:266-270 SKILL.md:270-283 SKILL.md:283-287 SKILL.md:287-301 SKILL.md:301-308 SKILL.md:308-313 SKILL.md:313-316 SKILL.md:316-320 SKILL.md:320-322 SKILL.md:322-334 SKILL.md:334-335 SKILL.md:335-336 SKILL.md:336-339 SKILL.md:339 SKILL.md:339-340 SKILL.md:340-341 SKILL.md:341 SKILL.md:341-344 SKILL.md:344 SKILL.md:344-345 SKILL.md:345 SKILL.md:345-348 SKILL.md:348-353 SKILL.md:353-354 SKILL.md:354-355 SKILL.md:355-356 SKILL.md:356-357 SKILL.md:357-358 SKILL.md:358-359 SKILL.md:359-360 SKILL.md:360-361 SKILL.md:361-362 SKILL.md:362-363 SKILL.md:363-364
🔑 環境變數 (1)
審計者: claude 查看審計歷史 →
品質評分
45
架構
100
可維護性
83
內容
20
社群
100
安全
91
規範符合性
你能建構什麼
建立預測模型
使用專門的時間序列演算法對感測器讀數進行分類、預測設備故障或預測銷售趨勢。
分析時間模式
偵測生理訊號中的異常、分割基因組資料，或在實驗序列中尋找重複的motifs。
擴展時間序列管線
與 sklearn 管線整合、比較基準測試結果，並部署可投入生產的時間序列模型。
試試這些提示
基本分類
Use aeon to classify the time series data in X_train, y_train. Train a RocketClassifier and evaluate accuracy on X_test.
異常偵測
Detect anomalies in my time series y using aeon. Use the STOMP detector with window_size=50 and return indices where anomaly_scores exceed the 95th percentile.
預測
Train an ARIMA forecaster (order=(1,1,1)) on my training data and predict the next 5 values.
管線優化
Create an sklearn pipeline with Normalizer, RocketTransformer, and a GradientBoostingClassifier using aeon transformers and sklearn estimators.
最佳實務
- 在使用大多數演算法之前，使用 aeon.transformations.collection.Normalizer 標準化時間序列資料
- 在嘗試深度學習方法之前，先使用 MiniRocketClassifier 或 RocketClassifier 進行快速原型開發
- 使用交叉驗證驗證模型效能，並與 1-NN Euclidean 等基準方法進行比較
避免
- 對樣本數少於 100 個的小型資料集使用深度學習分類器
- 在使用 DTW 或其他彈性距離度量時跳過資料標準化
- 忽視所需的輸入形狀 (n_samples, n_channels, n_timepoints)
常見問題
aeon 的輸入格式是什麼？
Aeon 期望時間序列的形狀為 (n_samples, n_channels, n_timepoints)。單變量資料的 n_channels 應為 1。
哪個分類器最快？
MiniRocketClassifier 提供最佳的速度效能平衡。RocketClassifier 也很快，且具有更多核心。
我可以使用的 sklearn 管線嗎？
是的，aeon 的轉換器和分類器與 sklearn Pipeline 和 GridSearchCV 完全相容。
如何偵測異常？
使用 aeon.anomaly_detection 中的 STOMP 或 LeftSTAMP。擬合後呼叫 fit_predict 來獲取異常分數。
有哪些距離度量可用？
DTW、DDTW、WDTW、Euclidean、Manhattan、LCSS、ERP、EDR、TWE、MSM 和 Shape-DTW。
如何處理缺失值？
在應用分類器或預測器之前，使用 aeon.transformations.collection 中的 SimpleImputer 處理缺失值。