人工智能Java SDK：kafka-情感倾向分析【英文】

最新推荐文章于 2024-09-12 02:53:14 发布

Calvin.AIAS

最新推荐文章于 2024-09-12 02:53:14 发布

阅读量1.4k

点赞数

分类专栏： AIAS 文章标签：人工智能 java kafka

本文链接：https://blog.csdn.net/weixin_39355136/article/details/121177391

版权

AIAS 专栏收录该内容

77 篇文章 20 订阅

订阅专栏

kafka-情感倾向分析【英文】SDK

情感倾向分析（Sentiment Classification）
针对带有主观描述的文本，可自动判断该文本的情感极性类别并给出相应的置信度，
能够帮助企业理解用户消费习惯、分析热点话题和危机舆情监控，为企业提供有利的决策支持。

SDK算法：

情感倾向分为两类
Negative (消极)
Positive (积极)

1. 启动 zookeeper:

zookeeper-server-start /usr/local/etc/kafka/zookeeper.properties

2. 启动 kafka:

kafka-server-start /usr/local/etc/kafka/server.properties

3. 创建 topic:

kafka-topics --create --zookeeper localhost:2181 --replication-factor 1 --partitions 1 --topic twitter-data

4. 下载并导入测试数据（src/test/resources目录下, twitter data.txt文件）

点击下载测试数据

kafka-console-producer --broker-list localhost:9092 --topic twitter-data < data.txt

5. 运行例子 - SentimentAnalysisExample

运行成功后，命令行应该看到下面的信息:

...
content: is alone downstairs...working 
prediction: [
	class: "Negative", probability: 0.98781
	class: "Positive", probability: 0.01218
]
content: I feel bad for doing it 
prediction: [
	class: "Negative", probability: 0.99725
	class: "Positive", probability: 0.00274
]
content: @RyanSeacrest is it just me, or she hates anoop. i mean seriously, she's kinda mean to him. 
prediction: [
	class: "Negative", probability: 0.99816
	class: "Positive", probability: 0.00183
]
...