Knowledge Map Construction Exploration-Knowledge atlas

First, the preface

The Knowledge Atlas (Knowledge Graph) suddenly becomes very popular, oneself also is assigned to study constructs the Knowledge Atlas, also saw a lot of material in the middle, here wants to make a small note, facilitates the later inspection and the memo. Second, what is the knowledge map.

The world is composed of entities, not strings (things, not strings).

The knowledge map is actually the previous semantic web, just a different name. A knowledge map is a network of relationships that connect all the different kinds of information (heterogeneous information). The knowledge map provides the ability to analyze problems from the perspective of "relationship".

The concept of a knowledge map was first proposed by Google in 2012 and is designed to optimize existing search engines. Unlike traditional search engines based on keyword search, knowledge maps can be used to better query complex information, understand user's intention from semantic level, and improve search quality. For example, in Google's search box to enter "Marie Curie" (Induced Marie Curie) This query, Google will be on the right to provide the details of Madame Curie, such as personal profile, place of birth, birth and death, and even some historical figures related to Madame Curie, such as Einstein, Pierre · Curie (Madame Curie's husband) and so on.

Third, the construction process

Four, my groping

The construction of the knowledge map is generally divided into four steps above, in which "knowledge storage" is a very important step, which is related to your next tool learning. Generally speaking, the Knowledge Atlas has two kinds of storage forms, one kind is the traditional RDF storage, the other is the graph database storage, here recommends the graph database, because the Google Knowledge Atlas uses is the graph database form, moreover the RDF form learning cost is higher, the difficulty of getting started.

Here is my groping process:
It's just starting with the form of RDF, after all, when we first contacted the Knowledge Atlas, we are sure to use the visual tools provided by Stanford University to build the protege, I also learned a bit, found it is too cumbersome, inefficient, and not suitable for a larger knowledge map of the construction, The main reason for the eventual abandonment is poor support for Chinese. Then I learned that Jena, a Java framework that provides APIs to handle RDF based ontology data, facilitates semi-automatic construction, and then uses a combination of Jena and virtuoso (a database that can store RDF, which is said to be stardog more well), However, the discovery of traditional RDF query needs to use SPARQL, learning cost is very high, complex and difficult to use.
Finally, using the NEO4J graph data and the way it provides cypher query statements, the advantages are obvious, learning and easy to start a lot, and because neo4j is also a more mainstream map data, the data rich a lot (Chinese have two good books: "Figure data" and "neo4j combat", easy to get started), NEO4J itself also provides the function of visualization data, which facilitates the construction of knowledge base.

The above is my groping process, only for everyone's reference, purely personal point of view.

If there are also in the exploration of the construction of knowledge map of the students or predecessors, welcome message exchange and guidance, I also continue to explore and learn. O (∩_∩) o