從文字擷取見解
使用自然語言處理 (NLP)，您無須具備任何機器學習專長，就能發掘非結構化文字的見解。識別關鍵片語和實體 (例如人物、地點和組織)，以了解共通的主題和趨勢。使用預先訓練之特定領域的模型來分類醫學術語。使用情感分析深入了解客戶的想法。評估各種語言的文字。
廣泛擷取實體
確認文字中的重要概念，包含關鍵片語與具名實體，例如人員、地點與組織。
強大的情感分析
了解客戶對您品牌的看法，並偵測有關特定主題的情感。
健全的語言偵測
評估各種語言的文字輸入。
彈性部署
以容器在任何位置 (雲端、內部部署或邊緣) 執行文字分析。
| 語言: | English (信心：100%) |
|---|---|
| 關鍵片語: | place, online menu, great menu, marvelous food, midtown NYC, week, dinner party, Contoso Steakhouse, pre-order, John Doe, Sirloin steak, chief cook, owner, kitchen, spot, dining, complaint, email |
| 人氣: |
文件
MIXED
86%
正面
0%
中立
14%
負面
例句 1
POSITIVE
99%
正面
1%
中立
0%
負面
例句 2
POSITIVE
100%
正面
0%
中立
0%
負面
例句 3
POSITIVE
100%
正面
0%
中立
0%
負面
例句 4
POSITIVE
100%
正面
0%
中立
0%
負面
例句 5
POSITIVE
100%
正面
0%
中立
0%
負面
例句 6
NEUTRAL
0%
正面
100%
中立
0%
負面
例句 7
NEGATIVE
0%
正面
0%
中立
100%
負面
例句 8
POSITIVE
100%
正面
0%
中立
0%
負面
|
| 具名實體: | Contoso Steakhouse [Location] midtown NYC [Location] last week [DateTime-DateRange] chief cook [PersonType] owner [PersonType] John Doe [Person] kitchen [Location-Structural] www.contososteakhouse.com [URL] 312-555-0176 [Phone Number] email [Skill] order@contososteakhouse.com [Email] |
| PII 實體: |
Type: Person Value: John Doe Type: URL Value: www.contososteakhouse.com Type: Phone Number Value: 312-555-0176 Type: Email Value: order@contososteakhouse.com |
| 連結實體: | We went to Contoso Steakhouse located at midtown NYC last week for a dinner party, and we adore the spot! They provide marvelous food and they have a great menu. The chief cook happens to be the owner (I think his name is John Doe) and he is super nice, coming out of the kitchen and greeted us all. We enjoyed very much dining in the place! The Sirloin steak I ordered was tender and juicy, and the place was impeccably clean. You can even pre-order from their online menu at www.contososteakhouse.com, call 312-555-0176 or send email to order@contososteakhouse.com! The only complaint I have is the food didn't come fast enough. Overall I highly recommend it! |
{
"languageDetection": {
"documents": [
{
"id": "153a065d-f501-488b-bad8-0423d54008bc",
"detectedLanguages": [
{
"name": "English",
"iso6391Name": "en",
"score": 1.0
}
]
}
],
"errors": [],
"modelversion": "2019-10-01"
},
"keyPhrases": {
"documents": [
{
"id": "153a065d-f501-488b-bad8-0423d54008bc",
"keyPhrases": [
"place",
"online menu",
"great menu",
"marvelous food",
"midtown NYC",
"week",
"dinner party",
"Contoso Steakhouse",
"pre-order",
"John Doe",
"Sirloin steak",
"chief cook",
"owner",
"kitchen",
"spot",
"dining",
"complaint",
"email"
]
}
],
"errors": [],
"modelversion": "2020-07-01"
},
"sentiment": {
"documents": [
{
"id": "153a065d-f501-488b-bad8-0423d54008bc",
"sentiment": "mixed",
"documentscores": {
"positive": 0.86,
"neutral": 0.0,
"negative": 0.14
},
"sentences": [
{
"sentiment": "positive",
"sentencescores": {
"positive": 0.99,
"neutral": 0.01,
"negative": 0.0
},
"offset": 0,
"length": 105
},
{
"sentiment": "positive",
"sentencescores": {
"positive": 1.0,
"neutral": 0.0,
"negative": 0.0
},
"offset": 106,
"length": 55
},
{
"sentiment": "positive",
"sentencescores": {
"positive": 1.0,
"neutral": 0.0,
"negative": 0.0
},
"offset": 162,
"length": 137
},
{
"sentiment": "positive",
"sentencescores": {
"positive": 1.0,
"neutral": 0.0,
"negative": 0.0
},
"offset": 300,
"length": 41
},
{
"sentiment": "positive",
"sentencescores": {
"positive": 1.0,
"neutral": 0.0,
"negative": 0.0
},
"offset": 342,
"length": 85
},
{
"sentiment": "neutral",
"sentencescores": {
"positive": 0.0,
"neutral": 1.0,
"negative": 0.0
},
"offset": 428,
"length": 139
},
{
"sentiment": "negative",
"sentencescores": {
"positive": 0.0,
"neutral": 0.0,
"negative": 1.0
},
"offset": 568,
"length": 62
},
{
"sentiment": "positive",
"sentencescores": {
"positive": 1.0,
"neutral": 0.0,
"negative": 0.0
},
"offset": 631,
"length": 30
}
]
}
],
"errors": [],
"modelversion": "2020-04-01"
},
"entities": {
"documents": [
{
"id": "153a065d-f501-488b-bad8-0423d54008bc",
"entities": [
{
"text": "Contoso Steakhouse",
"type": "Location",
"subtype": null,
"offset": 11,
"length": 18,
"score": 0.46
},
{
"text": "midtown NYC",
"type": "Location",
"subtype": null,
"offset": 41,
"length": 11,
"score": 0.55
},
{
"text": "last week",
"type": "DateTime",
"subtype": "DateRange",
"offset": 53,
"length": 9,
"score": 0.8
},
{
"text": "chief cook",
"type": "PersonType",
"subtype": null,
"offset": 166,
"length": 10,
"score": 0.82
},
{
"text": "owner",
"type": "PersonType",
"subtype": null,
"offset": 195,
"length": 5,
"score": 0.71
},
{
"text": "John Doe",
"type": "Person",
"subtype": null,
"offset": 222,
"length": 8,
"score": 0.94
},
{
"text": "kitchen",
"type": "Location",
"subtype": "Structural",
"offset": 272,
"length": 7,
"score": 0.6
},
{
"text": "www.contososteakhouse.com",
"type": "URL",
"subtype": null,
"offset": 477,
"length": 25,
"score": 0.8
},
{
"text": "312-555-0176",
"type": "Phone Number",
"subtype": null,
"offset": 509,
"length": 12,
"score": 0.8
},
{
"text": "email",
"type": "Skill",
"subtype": null,
"offset": 530,
"length": 5,
"score": 0.8
},
{
"text": "order@contososteakhouse.com",
"type": "Email",
"subtype": null,
"offset": 539,
"length": 27,
"score": 0.8
}
]
}
],
"errors": [],
"modelversion": "2020-04-01"
},
"entityLinking": {
"documents": [
{
"id": "153a065d-f501-488b-bad8-0423d54008bc",
"entities": [
{
"name": "Sirloin steak",
"matches": [
{
"text": "Sirloin steak",
"offset": 346,
"length": 13,
"score": 0.69
}
],
"language": "en",
"id": "Sirloin steak",
"url": "https://en.wikipedia.org/wiki/Sirloin_steak",
"datasource": "Wikipedia"
},
{
"name": "Steakhouse",
"matches": [
{
"text": "Steakhouse",
"offset": 19,
"length": 10,
"score": 0.78
}
],
"language": "en",
"id": "Steakhouse",
"url": "https://en.wikipedia.org/wiki/Steakhouse",
"datasource": "Wikipedia"
},
{
"name": "John Doe",
"matches": [
{
"text": "John Doe",
"offset": 222,
"length": 8,
"score": 0.01
}
],
"language": "en",
"id": "John Doe",
"url": "https://en.wikipedia.org/wiki/John_Doe",
"datasource": "Wikipedia"
},
{
"name": "New York City",
"matches": [
{
"text": "NYC",
"offset": 49,
"length": 3,
"score": 0.39
}
],
"language": "en",
"id": "New York City",
"url": "https://en.wikipedia.org/wiki/New_York_City",
"datasource": "Wikipedia"
}
]
}
],
"errors": [],
"modelversion": "2020-02-01"
},
"entityPII": {
"documents": [
{
"id": "153a065d-f501-488b-bad8-0423d54008bc",
"entities": [
{
"text": "John Doe",
"type": "Person",
"subtype": null,
"offset": 222,
"length": "8",
"score": 0.94
},
{
"text": "www.contososteakhouse.com",
"type": "URL",
"subtype": null,
"offset": 477,
"length": "25",
"score": 0.8
},
{
"text": "312-555-0176",
"type": "Phone Number",
"subtype": null,
"offset": 509,
"length": "12",
"score": 0.8
},
{
"text": "order@contososteakhouse.com",
"type": "Email",
"subtype": null,
"offset": 539,
"length": "27",
"score": 0.8
}
]
}
],
"errors": [],
"modelversion": "2020-07-01"
}
}
識別重要概念並分類
使用具名實體辨識，將文字中各種實體分類，例如人物、地點、組織、日期/時間和百分比。從文件中偵測及擷取 100 種以上的個人識別資訊 (PII)，包括受保護的健康狀態資訊 (PHI)。
擷取非結構化文字中的關鍵片語
快速評估及識別非結構化文字中的重點。使用關鍵片語擷取，取得最能描述每筆記錄主旨的相關片語清單。輕鬆提取及整理資訊，以了解重要主題和趨勢。
更準確地了解客戶的想法
偵測社交媒體、客戶評論及其他來源中的正面和負面情感，以探知對您品牌的想法。使用意見採礦，探索客戶對文字中各層面的感受，例如產品或服務的特定屬性。
處理非結構化醫療資料
利用文字分析的健康狀態功能 (目前為公開預覽)，從非結構化臨床文件 (例如醫囑、電子醫療記錄及病患諮詢表) 中擷取見解。辨識、分類及判斷醫療概念 (例如診斷、症狀，以及藥物的劑量與頻率) 之間的關聯性。
完整的隱私權與安全性
- 您的資料仍屬您所有。Microsoft 不會使用對您文字執行的訓練來改進模型。
- 選擇認知服務在哪裡使用容器處理您的資料。
- 文字分析由 Azure 基礎結構所支援，提供企業級安全性、可用性、合規性與管理性。
以彈性的定價，獲得您所需的功能、掌控度和自訂能力
- 用多少付多少，不用預付金。
- 使用文字分析時，只需依據交易數隨用隨付。
受到各種規模公司的信任
Wilson Allen 從非結構化資料發現見解
Wilson Allen 建立了功能強大的 AI 解決方案，可協助世界各地的法律與專業服務事務所，在原先互不相通且非結構化的資料中，發現前所未有的見解。
Kotak Mahindra Bank 加速提高生產力
Kotak Asset Management 採用聊天機器人為客戶服務管理進行轉型，以輕鬆地分析主旨列、客戶資訊和電子郵件內容，以識別情感並觸發下一個最佳動作。
文字分析 的常見問題集
-
文字分析可偵測各種語言、變體和方言。如需詳細資訊，請參閱語言支援文件。
-
不可以，模型都是預先訓練過的。您可以對上傳資料執行評分、關鍵片語擷取和語言偵測作業。若要建立和裝載自訂模型，請探索 Language Understanding 服務。
-
關鍵片語擷取會排除不必要的字詞和單獨的形容詞。形容詞與名詞詞組 (例如「壯觀的景象」或「多霧的天氣」) 會一起傳回。 一般而言，輸出由句子的名詞和受詞組成，並依重要性列出。重要性的測量依據為提到特定概念的次數，或文字中該項目與其他項目的關聯。
-
如果模型和演算法的改進內容屬於大幅變更，則會特別宣布，若只是微幅更新，則會直接新增至服務。過了一段時間，您可能會發現輸入相同的文字後，會輸出不同的情感分數或關鍵片語。在雲端中使用受控機器學習資源會有這種結果是正常的，也是刻意這麼做的。