Brief introduction of KNIME system
Brief introduction
KNIME (KonstanzInformationMiner, http://www.knime.org) system) is a data mining tool carefully developed based on Eclipse development environment. No need to install, easy to use. KNIME is also developed in Java and can be extended to use mining algorithms in Weka. Unlike YALE, KNIME uses a data flow-like (data flow) approach to establish an analysis mining process. The mining process consists of a series of functional nodes (node), each of which has an input / output port (port) for receiving data or models and exporting results. Each node in the KNIME has a traffic light to indicate the status of the node (red light when not connected, not configured, lack of input data; yellow light when ready for execution; green light after execution). There is a feature in KNIME-HiLite, which allows users to mark the records of interest in the node results and further explore.
Main function
The main features of KNIME include:
Using a fully graphical mode of operation, the following are the main operation interfaces of KNIME:
Support all kinds of data loading, including files, databases, etc.
Supports all kinds of data processing methods, including by column (such as split, merge, etc.), by row (filter, deformation), matrix (transpose) and PMML (field projection, one-to-many, many-to-one, normalization, anyway, etc.)
Support all kinds of data views, such as point chart, histogram, pie chart, distribution chart
Support hypothesis testing and regression methods
Support decision tree, Bayesian, clustering, rule derivation, neural network and other mining methods
Support process control
advantage
The KNIME system has the following advantages:
It provides a completely graphical mode of operation, the operation flow is simple, and the result output is intuitive.
Provides rich data reading and processing operations to support data acquisition from the database
Provide a more complete method of data mining
The report is rich and supports filling in different colors.
Deficiency
However, the KNIME system also has the following shortcomings:
Because it is mainly based on pattern operation, it is difficult to integrate with other systems.
two。 The support for statistical models is slightly insufficient.
