The Study Of Intelligence Judgment And Intelligence Forecasting Technology About Production Data Multi-dimensional Analysis System

The lift of electrolyze and the output of aluminum quantity are the extremely important factor which impact on the economic efficiency of enterprises. Therefore, the task of this paper is intelligent evaluation about the status of electrolyze and intelligent forecast about aluminum quantity.In the paper, I studied multi-dimensional analysis and data mining technology, and combined them as one, and combined with the characteristics of the data of aluminum electrolysis production, and carried on thorough theory and the experiment study. And designed and accomplished the aluminum electrolysis production data intelligent Judgment and intelligent forecast system. In this paper, content and innovation are as follows:1,I did the theoretical study of the data warehouse and multi-dimensional analysis and data mining technology. I proposed the architecture of intelligent evaluation and intelligent forecast system which base on multi-dimensional analysis and data mining technology. I used data warehouse and multi-dimensional analysis and data mining technology to collection electrolytic data which I needed. Then used the data mining technology to excavation to the data, and obtained rule which in the data contains. I used these rules to carry on to intelligent judgment the trough situation, to intelligent forecast the output of aluminum.2,In data mining pretreatment process, in view of the massive vacancy value, proposed the average value fluctuation interchange method. On the basis of the average value to add a coefficient whose range is between -N and +N to avoid the enormous influence which the insertion massive same values bring to decision tree's production.3,Proposed a ID3 decision-making algorithm which is in view of the aluminum electrolysis production data characteristic. The ID3 algorithm is one kind of inductive learning algorithm, it by the recursion defeat in detail way structure decision tree from the top, it can only deal with the logic of data or a discrete data, and ID3 algorithm production's tree node depend on the more number of features of data. In view of ID3 algorithm's characteristic, and unifies the characteristic of aluminum electrolysis production data, I proposed the improvement weighting ID3 algorithm and the processes of judgment and the forecast.