docs(ava/insight): provide instruction about Insight Extractors

Author: LAI-X

site/docs/api/insight/auto-insights.en.md

@@ -30,6 +30,20 @@ Run different algorithms from multi-dimensional data to discover interesting pat
 | algorithmParameter | `AlgorithmParameter` | Adjustable algorithm parameters | none |
 | dataProcessInfo | `Extra` | Configuration of data processing during data verification | none |
 
+* ***InsightType*** Supported Insight Types
+
+The extractor types used by Insight are set by the ***insightTypes*** property in the configuration. There are 7 extractor types in total, and by default, Insight will run each algorithm to provide the results.
+
+| Type | Name | Algorithm | Description |  
+| ----| ---- | ---- | ---- |
+| `trend ` | Trend Detection | [Mann-Kendall test](https://cran.r-project.org/web/packages/trend/vignettes/trend.pdf#page=2) | Detect significant increases or decreases in time series. | 
+| `change_point ` | Change Point Detection | [Bayesian Online Changepoint Detection](https://arxiv.org/pdf/0710.3742.pdf) | Detect change points in time series. | 
+| `majority ` | Majority Detection | proportion limits | Detect the outstanding indicators of data. | 
+| `low_variance ` | Low Variance Detection | [coefficient of variation](https://en.wikipedia.org/wiki/Coefficient_of_variation) | Detect the uniformity and determine if data points are distributed around the mean value. |
+|  `category_outlier ` | Category Outlier Detection | [IQR](https://en.wikipedia.org/wiki/IQR) | Detect category outliers in the data. | 
+| `time_series_outlier `  | Time Series Outlier Detection| [STL decomposition](https://otexts.com/fpp2/stl.html) | Detect outlier points in time series.  |  
+| ` correlation` | Correlation Detection | [Pearson product-moment correlation coefficient](https://en.wikipedia.org/wiki/Pearson_correlation_coefficient) | Detect if there is a correlation between two sequences of data. | 
+
 * ***AlgorithmParameter*** Adjustable algorithm parameters
 
 | Properties | Type | Description | Default|   

site/docs/api/insight/auto-insights.zh.md

@@ -33,6 +33,20 @@ order: 1
 | algorithmParameter | `AlgorithmParameter` | 可调的算法参数 | 无 |
 | dataProcessInfo | `Extra` | 数据校验时数据处理的配置 | 无 |
 
+* ***InsightType*** 支持的洞察类型
+
+算子类型共有7种，默认情况下会运行每一个算法来返回洞察结果，也可设置InsightOptions['insightTypes']参数指定参与计算的洞察类型。
+
+| 类型值 | 名称 | 算法 | 描述 |
+| ----| ---- | ---- | ---- | 
+| `trend ` | 趋势检测 | [Mann-Kendall 检验](https://cran.r-project.org/web/packages/trend/vignettes/trend.pdf#page=2) | 检测时间序列数据在某一指标上是否存在显著增加或者减少。 | 
+| `change_point ` | 突变点检测 | [贝叶斯在线变点检测](https://arxiv.org/pdf/0710.3742.pdf) | 检测在时间序列中某一属性发生的突变。 | 
+| `majority ` | 显著性检测 | 限制比例 | 检测数据分布的显著性，判断是否有突出的数据指标。 | 
+| `low_variance ` | 低方差检测 | [变异系数](https://zh.wikipedia.org/zh-hans/%E5%8F%98%E5%BC%82%E7%B3%BB%E6%95%B0) | 检测数据分布的均匀性，判断数据点是否分布在均值附近。 |
+|  `category_outlier ` | 类别异常检测 | [ IQR ](https://zh.wikipedia.org/wiki/%E5%9B%9B%E5%88%86%E4%BD%8D%E8%B7%9D) | 检测数据中的异常类别。 | 
+| `time_series_outlier ` | 时间序列异常检测 | [STL 分解](https://otexts.com/fpp2/stl.html) | 检测在时间序列中的异常点。 |  
+| ` correlation` | 相关性检测 | [Pearson 积矩相关检验](https://zh.wikipedia.org/wiki/%E7%9A%AE%E5%B0%94%E9%80%8A%E7%A7%AF%E7%9F%A9%E7%9B%B8%E5%85%B3%E7%B3%BB%E6%95%B0?wprov=srpw1_0) |检测两个序列之间是否存在相关关系。 |
+
 * ***AlgorithmParameter*** 可调的算法参数
 
 | 属性 | 类型 | 描述 | 默认值 |  

site/docs/guide/insight/intro.en.md

@@ -40,6 +40,36 @@ getInsights(data, {
 });
 ```
 
+Customize insight types by the  ***insightTypes*** attribute. Detailed parameters are described in the [getInsights API](../../api/insight/auto-insight.en.md).
+
+```ts
+import { getInsights } from '@antv/ava';
+
+getInsights(data, {
+  limit: 10,
+  insightTypes: ['trend', 'time_series_outlier']
+});
+```
+
+Customize impact measures and weights by setting up ***impactMeasures*** and  ***impactWeight***.
+
+```ts
+import { getInsights } from '@antv/ava';
+
+getInsights(data, {
+  limit: 10,
+  // set the measures of the impact score
+  impactMeasures: [
+    { fieldName: 'life_expect', method: 'COUNT' },
+    { fieldName: 'pop', method: 'COUNT' },
+    { fieldName: 'fertility', method: 'COUNT' },
+  ],
+  // adjust the calculation weight of the relevant factors (significance, impact) 
+  // in the calculation of the insight score.
+  impactWeight: 0.5,
+});
+```
+
 ### insightPatternsExtractor Usage
 
 If you only want to get insights of a specific type, `insightPatternsExtractor` will be your first choice. Detailed input and output parameters are described in the [insightPatternsExtractor API](../../api/insight/insight-patterns-extractor.en.md).

site/docs/guide/insight/intro.zh.md

@@ -41,6 +41,36 @@ getInsights(data, {
 });
 ```
 
+通过配置 ***insightTypes*** 属性指定计算的洞察类型，具体内容详见 [getInsights API](../../api/insight/auto-insights.zh.md)。
+
+```ts
+import { getInsights } from '@antv/ava';
+
+getInsights(data, {
+  limit: 30,
+  insightTypes: ['trend', 'time_series_outlier']
+});
+```
+
+使用 ***impactMeasures*** 和 ***impactWeight*** 属性来自定义影响力指标和权重。
+
+```ts
+import { getInsights } from '@antv/ava';
+
+getInsights(data, {
+  limit: 10,
+  // 自定义影响力（Impact）分数的计算指标
+  impactMeasures: [
+    { fieldName: 'life_expect', method: 'COUNT' },
+    { fieldName: 'pop', method: 'COUNT' },
+    { fieldName: 'fertility', method: 'COUNT' },
+  ],
+  // 自定义影响力（Impact）分数在洞察分数中的权重（0 ~ 1）
+  impactWeight: 0.5,
+});
+```
+
+
 ### insightPatternsExtractor 使用
 
 如果只想获取指定类型的洞察结果，那么`insightPatternsExtractor`将是你的首选。输入输出参数详见 [insightPatternsExtractor API](../../api/insight/insight-patterns-extractor.zh.md)。