1.7 An Overview of User Insights and Clustering

Loading...

来自 University of Illinois at Urbana-Champaign 的课程

Cluster Analysis in Data Mining

125 个评分

University of Illinois at Urbana-Champaign

Cluster Analysis in Data Mining

125 个评分

课程 5（共 6 门，Specialization Data Mining ）

Discover the basic concepts of cluster analysis, and then study a set of typical clustering methodologies, algorithms, and applications. This includes partitioning methods such as k-means, hierarchical methods such as BIRCH, and density-based methods such as DBSCAN/OPTICS. Moreover, learn methods for clustering validation and evaluation of clustering quality. Finally, see examples of cluster analysis in applications.

从本节课中

Module 1

1.1. What is Cluster Analysis2:16

1.2. Applications of Cluster Analysis2:17

1.3 Requirements and Challenges5:12

1.4 A Multi-Dimensional Categorization2:19

1.5 An Overview of Typical Clustering Methodologies6:56

1.6 An Overview of Clustering Different Types of Data6:56

1.7 An Overview of User Insights and Clustering3:17

与讲师见面

Jiawei Han
Abel Bliss Professor
Department of Computer Science

[SOUND] Hi, in this session,

I'm going to give you a brief overview of user insights and clustering.

Usually, users may have some interesting insights or views.

User may like to interact with the clustering process,

to influence the final results of clustering.

One interesting one is visual insights,

because a picture is worth a thousand words.

Human eyes are actually powerful, high-speed processors.

Because they are also linked with rich knowledge base.

So, a human can provide a lot of intuitive insights

whether this clustering result is good or

something they desired is not achieved based on the current clustering result.

Okay, there's usually user's good to one or two dimensional,

at most 3D, they may be able to find clusters.

Their algorithms, for example, High D eyes,

how to visualize high dimensional clusters.

The second one is semi-supervised insights,

that means user may have certain insights or intention on the clustering process.

They may like to pass their insights to the system to

influence the results of the clustering.

For example, user may provide some seeding.

That means you can provide a number of labeled examples or

some approximate representation of the categories of interest.

Then you may like the clustering process to

take your insight to do some desired clustering.

The third one is multi-view and ensemble-based insights.

So-called multi-view clustering is different clustering

may generate different perspectives.

The multiple clustering,

you can ensemble them together to provide more robust solution.

Finally, a very important one is validation-based insight.

That means, you may want to evaluate the quality

of the clusters generated based on certain validation methods.

For example, you may use case studies or you may provide specific measures,

or you may provide some preexisting labels.

All of these, we're going to discuss in detail in some chapter.

Finally, I provide a few reference books.

And especially the chapter By Charu Aggarwal.

An Introduction to Clustering Analysis is a very

nice introduction on general ideas of clustering analysis.

Thank you.

[MUSIC]

探索我们的目录

免费加入并获得个性化推荐、更新和优惠。

Coursera 致力于普及全世界最好的教育，它与全球一流大学和机构合作提供在线课程。

© 2018 Coursera Inc. 保留所有权利。

通过 App Store 下载

通过 Google Play 获取