Precise and Accurate Job Cycle Time Forecasting in a Wafer Fabrication Factory with a Fuzzy Data Mining Approach

Table 3

Advantages and disadvantages of various data mining approaches for job cycle time forecasting.

Approach	Advantages	Disadvantages
Clustering	(i) Easiest to use; (ii) Can be used with other methods	(i) Not accurate; (ii) No method to determine the optimal number of clusters
GA	(i) Accurate; (ii) Can be used with other methods	Time-consuming iterations
Decision trees	(i) Handles missing values well; (ii) Robust to outliers	High variance of tree structure
ANN	(i) Accurate; (ii) Robust to outliers	(i) Difficult to use; (ii) Time-consuming training; (iii) Possible overfitting
Hybrid approaches	Very accurate	(i) Very difficult to use; (ii) Very time consuming