Data Mining And Application Options Plant

Data mining technology is a method of discovering potential and invisible new information from large amount of data. And new information is often utilized to conduct human activities. Initially, the data mining technology is used in commercial areas, such as forecasting customers or product sale quantity. Data mining includes data preparing, data mining processing and knowledge discovering, all these procedures are essential. Before data mining, all the data in a set should be preprocessed, which includes stages of data cleaning, data integration and data reduction. In the stage of data integration, all data from different data base are integrated into a single data set. And the knowledge discovering process in data contains many repetitions and transitions procedures.In order to discover the knowledge in data set efficiently, amount of methods and algorithms have been investigated. For instance, characterization, generalization, classification, cluster analysis, associate, evolution, pattern match etc. Characterization is used for summarizing the general characteristics of a dataset. However, classification is used for determining and classifying the class in a training set. Decision trees, regression analysis, artificial neutral networks, support vector machines and k-nearest neighbor algorithm are usual methods which widely utilized in classification. Cluster analysis is used for clustering similar data structures in any dataset. Cluster analysis is used for clustering similar data structures in any dataset. And hierarchical methods, partitioning methods, density-based methods are among the cluster analysis techniques.The associate analysis discover relationships of the processing dataset, and usually be used to solve two problems:(1) Finding frequent itemsets with their supports above the minimum support threshold;(2) Using frequent itemsets found in the step (1) to generate associate rules that have a confidence level above the minimum confidence threshold. Therefore, many algorithms are developed in the core of finding frequent itemset. Apriori algorithm is the earliest and most classical method utilized in associate rules analysis. Then many algorithms were researched in order to increase efficiency of algorithms, except Apriori-like algorithms, there have many other algorithms, such as:(1) FP-growth algorithm, which increase efficiency by scaning the dataset twice instead of generating candidate itemsets;(2) Parallel mining, which contains multiple processors in the computing environment, and the mining tasks are separated into several sub-tasks so that each sub-task can be performed on various processors.(3)Sampling algorithm, it’s shortcoming is there might exist some missing associate rules, and in this case, the approach have to scan two passes over the data.This paper researched the processing data from a multi-metal beneficiation plant, utilizing the data mining technology to analysis it’s processing flowsheet, regime of reagents, flotation indicators and so on. According to the research results:(1)Copper minerals, lead minerals and zinc minerals present a close intergrowth relationship, and it is essential to research its mineralogy in detail;(2)In copper flotation process, the dosage of lime, W-23,Y-541, zinc sulfate, sodium sulfite and sodium sulfide are over used, and need to adjust;(3)In zinc flotation process, both copper sulfate and Y-541dosage is feasible, but lime dosage is a little excess;(4)Reagents utilized in copper flotation have a great effect on zinc flotation;(5)There is a contradiction between copper flotation reagents regime and zinc flotation reagents regime. A choice should be made among copper recovery and zinc recovery, or find a reasonable reagents regime which will obtain a satisfying copper recovery and also zinc recovery.