Social Mining with PersonEngine
The web presents us with the largest accumulation of knowledge and information within recorded history. The mass and complexity of this information makes it very difficult to accurately model what is happening in a way that allows us to anticipate the results of publishing new content. Early attempts at managing this complexity rely on “authority” measures such as page rank to filter out information leaving a more reasonably sized subset to analyze. While efficient and useful for displaying relevant results to users this authority measure does nothing to help us understand features of a publication’s content that enhance its effectiveness in capturing the attention of web users. We are still faced with a large degree of uncertainty when trying to define what content will be successful.
Technology solutions for richer and more meaningful analysis of text content exist, unfortunately the solutions themselves are nearly as complex and overwhelming as the web itself. This has led to an environment where tools for things like analysis of sentiment, semantics or simply word usage on large scales are unavailable to anyone without expertise in machine learning or the means to hire it. Fortunately advances in human computer interaction have provided techniques for making complex expert systems accessible and understandable for those with less than a graduate level computer science education.
via Social Mining – learning the web with AI « PersonEngine.
Categories: Social Optimization