Postprocessing is like data preparation one of the most challenging tasks in data mining that users must deal with. It is desirable to simplify it so that the path to results is as fast and efficient as possible. The extension of the EasyMiner research project to filter the association rules by similarity to the knowledge-based rules, should be helped in this respect, which is the subject of this diploma thesis. The objectives were accomplished by a detailed analysis of the default state of EasyMiner in conjunction with a thoroughly thought-out implementation proposal without increasing the demands on the server or user of the application. An analysis of general practices and the author's deep knowledge of Internet application issues served to do this. The future deployment of this extension to the EasyMiner infrastructure will benefit from a clearer and more efficient work with the Knowledge base part, where it will no longer be necessary to evaluate the interest in the same or similar rules and the user will be able to focus directly on the quality of the results. This thesis is divided into chapters as a detailed description of how a similar problem can be approached in any other project that works with a certain form of knowledge base. Initial input analysis with access search for comparison of the various elements passes through a description of the default application state to a specific solution design. This should be a guideline for the implementation itself and for the testing of the proposed and implemented procedures.
| Identifer | oai:union.ndltd.org:nusl.cz/oai:invenio.nusl.cz:358853 |
| Date | January 2017 |
| Creators | Duben, Přemysl Václav |
| Contributors | Vojíř, Stanislav, Zeman, Václav |
| Publisher | Vysoká škola ekonomická v Praze |
| Source Sets | Czech ETDs |
| Language | Czech |
| Detected Language | English |
| Type | info:eu-repo/semantics/masterThesis |
| Rights | info:eu-repo/semantics/restrictedAccess |
Page generated in 0.002 seconds