Abstract
The SECOM dataset contains information about a semiconductor production line, entailing the products that failed the in-house test line and their attributes. This dataset, similar to most semiconductor manufacturing data, contains missing values, imbalanced classes, and noisy features. In this work, the challenges of this dataset are met and many different approaches for classification are evaluated to perform fault diagnosis. We present an experimental evaluation that examines 288 combinations of different approaches involving data pruning, data imputation, feature selection, and classification methods, to find the suitable approaches for this task. Furthermore, a novel data imputation approach, namely “In-painting KNN-Imputation” is introduced and is shown to outperform the common data imputation technique. The results show the capability of each classifier, feature selection method, data generation method, and data imputation technique, with a full analysis of their respective parameter optimizations.
Highlights
A semiconductor manufacturing process holds a colossal amount of data that are gathered from many sensors during the process
The data from the SECOM dataset goes through the stages of data pruning, data imputation, feature selection, and data boost and is classified using ten-fold cross-validation
True negative ratio (TNR) indicates the percentage of successes that were rightly predicted and indicated how few false alarms were given by the classifier
Summary
A semiconductor manufacturing process holds a colossal amount of data that are gathered from many sensors during the process. The gathered data contains valuable information regarding each produced entity, which can be used to gain knowledge about the yield of the production line, the failures that occurred and their results, and many other information that are crucial in operation of the semiconductor plant. How this data is handled is of great importance. The incorporating stages in processing these wafers can include: (1) Cleaning the wafer, in which the base of the semiconductor is cleaned To accomplish this task, chemical substances are employed that can remove minor particles and residues made in the process of production or created by the exposure to air. The wafer is rotated by centrifugal force with the aim of creating a uniform layer of resistance on the wafer surface. (5) Exposure, in which the wafer is exposed to short wavelength deep ultraviolet radiations through a mask (containing the formed design patterns). (6) Development, in which the wafer is sprayed with the purpose of dissolving the exposed areas and revealing the thin film on the wafer surface. (7) Etching, in which a pattern is created on the wafer by either: (a) excessive attacking on the wafer by ionized atoms for elimination of the film layer; (b) dissolving and removing the surface layer using chemicals (such as hydrofluoric acid or phosphoric acid). (8) Insertion of impurities, in which certain elements such as phosphor or boron ions are inserted into the wafer, creating the semiconducting properties. (9) Activation, in which the substrate is heated quickly via certain tools such as flash lamps. (10) Assembly, in which the output wafer is cleaned before it is separated into individual chips through dicing. (11) Packaging
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
