Glossary

Analytics (statistical inferencing) Chatmine uses a variety of statistical techniques to provide you inferential testing. We use factor, PCA, decision or probability trees, discriminant analysis and test the predictions (T, F and Chisquare tests). These give you a deeper understanding as well as relevance of our findings to larger populations. For example, is there a significant difference between gaming purchases made between expert and novice players?

Blog (Weblog) is a web-based publication consisting primarily of periodic articles (normally in reverse chronological order) posted by individuals or groups. Chatmine identifies relevant blogs in your market with proprietary metrics.

Data Mining is the automatic searching of large databases for patterns using computational algorithms or statistical techniques. Chatmine employs data mining on the datasets it collects from the web.

Decision Tree is a predictive model that is a mapping of observations about an item to conclusions about the item's target value. Each inner node corresponds to variable; an arc to a child represents a possible value of that variable.

Domain Expertise is a high level of knowledge about a particular category or topic (domain). A linguist is said to have domain expertise in linguistics, a physician in medicine, etc.

Factor analysis a statistical approach that can be used to analyze interrelationships among a large number of variables and to explain these variables in terms of their common underlying dimensions (factors).

Hierarchical Data Base is an older form of database usually found on mainframes and also used for bulletin boards and online communities. Hierarchical databases are simple but inflexible, because all of the data must travel with a given name or a single identifier; e.g., as discussion threads travel with a topic.

Information Visualization is the translation of data into understandable schemas, analogies and metaphors. Here are just a few of the conceptual and visual tools Chatmine uses:

semantic network: maps the relationships between concepts in a spider figure; deep data: presents relationships between concepts in multilayered tiled representation; social network analysis: helps identify thought/opinion leaders and cliques of consumers as well as external influencers, outliers—visualizes trends, statistically significant relationships and clusters in the data easily.

Latent semantic analysis provides the frequency with which each term occurs in each document out of a collection of documents. Chatmine uses this technique to map your consumer perceptual and decision space.

Lexicon is a set of key terms/concepts that a particular community uses to express itself concerning the things it cares about. Thus a video gaming community might talk about gameplay, graphics, physics, multiplayer function, cheat codes, patches, etc.

Pattern recognition or pattern classification enables organization of raw data into meaningful categories.
Supervised: an expert's domain expertise can be incorporated into the classification process.
Unsupervised: works better for new and novel datasets, where there is little or no expertise.

Semantic networks A semantic network is often used as a form of knowledge representation. It consists of nodes which represent concepts and edges which represent semantic relations between the concepts. Using artificial intelligence (AI) and pattern recognition algorithms, conversations, discussions and chats are mined and concepts are extracted. Relationships between the concepts (nodes) are form links in the networks. The strength of nodes and relationships are measured providing a means of understanding the consumer decision space.

Social networks A social network is a map of the relationships between individuals, ranging from casual acquaintance to close familial bonds. Virtual communities are built around affinity and similarity. Chatmine lets you monitor the discussions taking place around the "virtual watercoolers." Our quantitative analysis lets you identify the thought and opinion leaders of community, size and shape of communities as well as external influencers.

Sparse data set is an incomplete set of data that makes the drawing of meaningful inferences problematic. These kind of data are often found in answers to open-ended survey questions. One way to make such data more meaningful is to identify patterns of relationships in the sparse data set and match those patterns to like patterns in larger data sets (e.g., online communities).