Open In App

What is Data Mining Trends and Research Frontiers?

Improve
Improve
Like Article
Like
Save
Share
Report

Data mining is the process of analyzing a large size of information to find out the patterns, trends. It can be used for corporations to find out about customers’ choices, make a good relationship with customers, increase the revenue, reduce risks. Data mining is based on complex algorithms that allow data segmentation to discover numerous trends and patterns, detect deviations, and estimate the likelihood of certain occurrences occurring. Raw data can be in both analog and digital formats, and it is essentially dependent on the data’s source. Companies must keep up with the latest data mining trends and stay current in order to succeed in the industry and beat out the competition.

Types of Mining Sequence in Data Mining:

  • Mining time series
  • Mining symbolic sequence
  • Mining biological sequence

1. Mining Time Series

A specified number of data points are recorded at a specific time or events obtained over repeated measurements of time in a mining time series. The values or data are typically measured in equal time intervals like- hourly, weekly, daily. In time-series data is also recorded regular intervals or characteristic time-series components are trend, seasonal, cycle, irregular.

Application of Time Series:

  • Financial: Stock market analysis
  • Industry: Power consumption
  • Scientific: Experiment result
  • Meteorological: Precipitation

Time Series Analysis Methods:

  • Trend Analysis: Categories of Time Series movements:
    • Long-term or Trend Movements: General direction in which a time series is moving over a long interval of time.
    • Cyclic Movements: Long-term oscillation about a trend line or curve.
    • Seasonal Movements: A time series appears to follow substantially identical patterns during the corresponding months of subsequent years.
    • Irregular or Random Movements: It changes that occur randomly due to unplanned events.
       
  • Similarity Search:
    • Data Reduction
    • Indexing Methods
    • Similarity Search Methods
    • Query Languages

2. Mining Symbolic Sequence

A symbolic sequence is made up of an ordered list of elements that can be recorded with or without a sense of time. This sequence can be used in a variety of ways, including consumer shopping sequences, web clickstreams, software execution sequences, biological sequences, and so on.

Mining of sequential patterns entails identifying the subsequences that appear frequently in one or more sequences. As a result of substantial research in this area, a number of scalable algorithms have been developed. Alternatively, we can only mine the set of closed sequential patterns, where a sequential pattern s is closed if it is a correct subsequence of s’ and s’ has the same support as s.

For example:

if S = [ { ab }, d ]   S' = [ { abc }, { be }, { de }, a ],    where a, b, c, d and e are items, then S is a subsequence of S’.

3. Mining Biological Sequence

Biological sequences are made up of nucleotide or amino acid sequences. In bioinformatics and modern biology, biological sequence analysis compares, aligns, indexes, and analyzes biological sequences. Biological sequences analysis plays a crucial role in bioinformatics and modern biology. Such analysis can be partitioned into two tasks- pairwise sequence alignment and multiple sequence alignment. 

Biological Sequence Methods:

  • Alignment of Biological Sequences:
    • Pairwise Alignment
    • The BLAST Local Alignment Algorithm
    • Multiple Sequence Alignment Methods
  • Biological Sequence Analysis Using a Hidden Markov Model:
    • Markov Chain
    • Hidden Markov Model
    • Forward Algorithm
    • Viterbi Algorithm
    • Baum-Welch Algorithm

Application of Data Mining: 

  • Financial Information Analysis:
    • Loan payment prediction/consumer credit policy analysis
    • Design and construction of information warehouse
    • Financial information collected in banks and money establishments area units are typically comparatively complete, reliable, and of top quality.
  • Retail Industry:
    • Multidimensional analysis( sales, customers, products, time, etc.)
    • Sales campaign analysis
    • Customer retention
    • Product recommendation
    • Using visualization tools for data analysis
  • Science and Engineering:
    • Data processing and data warehouse
    • Mining complex data types
    • Network-based mining
    • Graph-based mining

Trends of Data Mining: 

  • Exploration of applications: addressing application-specific issues
  • Data mining approaches that are scalable and interactive
  • Data mining integration with Web search engines, database systems, data warehouse systems, and cloud computing systems
  • Mining social and information networks
  • Mining spatiotemporal, moving objects, and cyber-physical systems
  • Mining multimedia, text, and web data
  • Mining biological and biomedical data
  • Visual and audio data mining
  • Distributed data mining and real-time data stream mining.

Last Updated : 10 Feb, 2022
Like Article
Save Article
Previous
Next
Share your thoughts in the comments
Similar Reads