## Data Mining

In industry lots of data available in business,science or any type of industry. Firstly that all data and daily transaction saved in operational database.in that operation database all data saved related with day to day transaction.

Data warehouse collect data from operational data warehouse and save successfully.In data warehouse gives only important data from operational database.if operational database contains 100 transaction then in data warehouse gives a 95 transactions from operational database.

Data mining basically coming from KDD (knowledge discovery database) concept.Data mining is only part of KDD process.

Data mining used from selecting data from data warehouse and show that data to user with with graphical formation like pi chart,bar chart ,diagram etc. Data mining select a important data from data warehouse with user requirement and show that data to user with graphical representation.

Data mining applications like R software and WEKA software are used in big industry for sorting data from big data and result data used for giving better decisions. The result data in the form of graphical manner and in the form of diagrammatic manner, so result is easy to understand. Data mining uses GUI visualization for result creation in the form of-

**Graphical-**pi chat,histogram,line graph etc**Geometric-**Include black spot and scatter diagram technique.**Icon Based-**Include figure,color etc**Pixel Based-**Colored picture.**Hierarchical-**This technique divides the display region based on data value.**Hybrid-**Combine many pictures in to one.

### Basic Data Mining Task-

In data mining lots of methods are used for getting data from data warehousing.Data mining basically divide into two parts for successfully getting data from data warehouse.

- Predictive
- Descriptive

These two types of important for data mining task.In data mining task that 2 types are again divide into mainly 4 types.

- Predictive
- Classification
- Prediction
- Regression
- Time Series Analysis

- Descriptive
- Clustering
- Summarization
- Association rule
- Sequence Discovery

Data mining uses that 8 methods for data distribution and collection of data from data warehouse.

#### A. Predictive-

In Predictive analysis allow user to predict future outcomes from current available data. In predictive analysis calculate future outcomes from present situation and from present data mining data.

In that only thinking about present or early information for calculate future outcomes. Predictive outcomes provide users with advice manner for what action taking in future for business.

Prediction divide in 4 types-

- Classification-

In classification method, mainly class name and object name are known to user for predict a value. eg.-car sale calculate in previous year so car name is class name and car color is indicating object name. we find in previous year total for x car and y color. In that class name and object name is known then is called as supervised classification.

2. Regression-

That method is only used for calculating numerical formations value. In data mining tool we need to calculate age,height or any numerical value data then that time regression method are used. In that only numerical values are identify for future outcomes.

3. Predictive-

If we want to calculate prediction value then use prediction method. In that factor current data and some historical facts are used to predict a values.

4. Time Series Analysis-

For calculating future prediction value time series analysis is used. In that formation used a time period information or time slot information. Information given in the form of time slot mainly for predict future outcomes.

#### B. Descriptive-

In descriptive manner also work for predict a future outcomes but in descriptive use one second ago data to previous all data for calculating predictive value. All information in the form of summary or in the form of description. Industry related all information from establishment is used as summery in that part and calculate future outcomes.

In that 4 types are described mainly.

- Clustering-

In that the class name is unknown and the object name is known. The objects are in form of sub classes. The sub classes are grouped that called cluster. In that class name is unknown that reason cluster is called as unsupervised classification.

2. Summarization-

In that only gives simple description of data. The summery about that data coming in that type. The characterization or generalization of that data are described in that class mainly. use of characteristics of data find future outcomes.

3. Association Rule-

In association rule thinking about possibilities between If-Then condition about any product and find outcomes for future. If two product sale with if-then condition sequentially then association rule apply on that respectively.

4. Sequence Discovery-

In that type sequence manner important. that depend upon time sequence of actions. if product sale sequencing in order then that time for finding future outcomes use that method with association rule.

Very knowledgeable information. Keep it up.