Professional Documents
Culture Documents
INDIVIDUAL ASSIGNMENT 1
CONSTRUCTION ISSUES: USING DATA MINING TECHNIQUES FOR IMPROVING BUILDING LIFE CYCLE
Contents
Definition ................................................................................................................................... 1 Application of data mining process in construction issue ......................................................... 1 For the first data mining technique: Visual analysis approach (Stacked Histogram) ................ 2 For the second data mining technique: Clustering Algorithm .................................................. 2 For the third data mining technique: Classification Tree Algorithm ......................................... 2 For the fourth data mining technique: Association Rule .......................................................... 3 References ................................................................................................................................. 3
Definition
Basically, data mining is about a process that analysis our data from many perspective then sum up those information to convert into useful information. We can study the knowledge through the data we have collected. the information we get will help in many fields that relate using database for examples hospital, construction project and weather forecasting. Data mining software is one of analytical tools for analyzing data. It allows users to analyze data from many different dimensions or angles, categorize it, and summarize the relationships identified. Besides that, data mining tools allow us to predict the behaviours and future trend, decision making for administrative department and knowledge- driven decisions. They are two models of data mining which are predictive and descriptive. In predictive model consist of classification, regression, time series analysis and prediction. Then, descriptive model will have clustering, summarization, association rules and sequence discovery.
For the first data mining technique: Visual analysis approach (Stacked Histogram)
A histogram is defined as a bar graph that shows frequency data. In a histogram, data is collected and sorted into categories. Analysis using histogram is a powerful technique for looking at processing large amount of data. As usual from this article, I notice that the interactive stacked histogram has helped to solve with the problems\ of incapability of cross comparison of standard histogram to allow different trend is analyzed under the same dynamic graph. This is the way to differentiate the correlation between attribute priority and cause-of-repair that can be visualized.
Clustering techniques are applied when there is no class to be predicted but rather when the instances are to be divided into natural groups (Witten and Frank, 2000). Clustering method is good in order to generate similar collections that simplify the representation of data sets. Actually, simplification data is very useful especially those data with very large scale that consist of multi-dimensional attributes. But, normally the clustering algorithm technique is use to identify the critical attributes in multidimensional space. In the article, the researchers using this method to cluster the industrial data into two clusters which centered with two major type of machine malfunction. The result they have obtained is considered as a potential knowledge for the maintenance and building management purpose in future. Apart for that, it also reduce the time consuming of the researchers as clustering algorithm techniques has simplified those massive data into cluster through clustering process.
A decision tree is a tree-based knowledge representation methodology used to present classification rules. In the article, the researcher using the classification tree algorithm to classify the monthly priority maintenance works were carried out in the later part of the year, July to November. All 6 monthly maintenance works happen to be in
December. As I noticed, the researchers use this method to know which maintenance works to be given priority first and fix the staring date of the maintenance works.
The association rule technique involves finding frequent patterns, associations, correlations, or casual structures among sets of items or objects in transaction databases, relational databases, and other information repositories (Han, 2001). Basically, the researchers have admitted this method is very efficient in searching the associations and correlations between attributes. But the researchers knew that to avoid having a large number of rules which do not bring any meaning toward the research, they have filtered out all the irrelevant attributes and find the groups of correlated attributes prior to applying the algorithm. The researchers have gone through the filtering process on the available maintenance data of the Air Handling Units Building. In the end, they found out that the Numeric data, string data types are not applicable to most of the associative rule algorithms.
References
Alexander, D. (n.d.) Data Mining. Retrieved 3 March 2014 from http:// www.laits.utexas.edu/~anorman/BUS.FOR/course.mat/Alex/
Arditi, D. and gunaydin, M.h. (1998) Factors that affect process queslity in the life cycle of building projects, Journal of Construction Engineering and Management,ASCE, 124 (3) : 194-203
Gero, S., Reffat, R. M., Peng, W., Liew, P.& Rosenblatt, J. (2003). Using data mining techniques for improving building life cycle. Retrieved 2 March 2014 from http://www.constructioninnovation.info/images/pdfs/Research_library/Researc hLibraryB/ProjectReports/200103_R_Using_Data_Mining_Techniques_for_i mproving_building_life_cycle.pdf