Question

We are interested in detecting communities in a social media dataset.
1. How would you mine this problem?
2. Choose a social network and explain the kind of data cleaning you need.
3. What kind of data mining algorithm can you use?

Answer 1

1. Recently, social networking sites are offering a rich resource of heterogeneous data. The analysis of such data can lead to the discovery of unknown information and relations in these networks. The detection of communities including ‘similar’ nodes is a challenging topic in the analysis of social network data, and it has been widely studied in the social networking community in the context of underlying graph structure. Online social networks, in addition to having graph structures, include effective user information within networks. Using this information leads to enhance quality of community discovery. In this study, a method of community discovery is provided. Besides communication among nodes to improve the quality of the discovered communities, content information is used as well. This is a new approach based on frequent patterns and the actions of users on networks, particularly social networking sites where users carry out their preferred activities. The main contributions of proposed method are twofold: First, based on the interests and activities of users on networks, some small communities of similar users are discovered, and then by using social relations, the discovered communities are extended. The F-measure is used to evaluate the results of two real-world data set demonstrating that the proposed method principals to improve the community detection quality.

2. If you are analyzing your social media activities without cleaning your data first, you are wasting your time. Clean data is the key to getting your strategy right.

In social media, conversations aren't always that clean.

By unclean conversations we mean irrelevant ones. With the huge amount of existing social conversations, there naturally comes a lot of irrelevant ones too.

When speaking of companies and social data, irrelevant data refers to spam, ads, posts by the company itself or its employees, as well as posts not related to the brand. In other words, "noise". Also, not everything is posted by real humans. Some can be created by social media robots, so called bots.

Let's say that ACME brand has received 12 500 messages during the past seven days. Of all the messages, 77 % was noise. Which means, only 23 % of the whole conversation is relevant and created by real humans. Hence, you can ask yourself, would you rather make your decisions based on spam, as in the total of 12 500 messages, or on those 2875 messages actually created by real users?

Filtering out and removing the noise is indeed very time consuming, yet vital. In marketing, as well as in product development, it’s crucial to base your decisions on reliable and relevant data in order to understand what your customers really want.

3. The Balanced Link Density Label Propagation Algorithm is  used .

The proposed method (BLDLP) substitutes random selection from famous LPA algorithm with a rational choice.

The BLDLP algorithm results are more stable

The BLDLP algorithm increases the original time efficiency