Datasetpublic

ofwestern/Glass_Classification

Glass Identification Database About This File This repository contains information about the Glass Identification Database, a well-known dataset used primarily for classification tasks in forensic and criminological investigations. Sources Creator:B. GermanCentral Research EstablishmentHome Office Forensic Science ServiceAldermaston, Reading, Berkshire RG7 4PN Technical note number: Unknown / not listed General results: Nearest Neighbor… See the full description on the dataset page: https://huggingface.co/datasets/ofwestern/Glass_Classification.

sourceHugging Facemitupdated 9mo agoView on Hugging Face

0likes1downloads

Dataset Card

Glass Identification Database

About This File

This repository contains information about the Glass Identification Database, a well-known dataset used primarily for classification tasks in forensic and criminological investigations.

Sources

Creator: B. German Central Research Establishment Home Office Forensic Science Service Aldermaston, Reading, Berkshire RG7 4PN

—Technical note number: Unknown / not listed
—General results: Nearest Neighbor performed competitively when compared to the rule-based system.

Relevant Information

Vina conducted a comparative study involving:

—BEAGLE (a rule-based system),
—Nearest Neighbor (NN) algorithm,
—Discriminant Analysis (DA).

BEAGLE is a commercial product available from: VRS Consulting, Inc. 4676 Admiralty Way, Suite 206 Marina Del Rey, CA 90292 Phone: (213) 827-7890 Fax: (213) 827-3189

Classification Results

Determining whether glass was float processed or not (number of incorrect answers):

Type of Sample	BEAGLE	NN	DA
Windows (float processed, 87)	10	12	21
Windows (non-float processed, 76)	19	16	22

The motivation behind this study is criminological investigation. Glass fragments left at crime scenes can serve as valuable evidence—if correctly identified.

Dataset Information

—Number of Instances: 214
—Number of Attributes: 10 (including an ID) + 1 class attribute
—Attribute Type: Continuous
—Missing Values: None

Attribute Information

Attribute	Description
Id	Identification number (1–214)
RI	Refractive Index
Na	Sodium (weight % in oxide)
Mg	Magnesium
Al	Aluminum
Si	Silicon
K	Potassium
Ca	Calcium
Ba	Barium
Fe	Iron
Type	Type of glass (class attribute)

Glass Type Classes

1.Building windows (float processed)
2.Building windows (non-float processed)
3.Vehicle windows (float processed)
4.Vehicle windows (non-float processed) (none in this dataset)
5.Containers
6.Tableware
7.Headlamps

Summary Statistics

Attribute	Min	Max	Mean	SD	Correlation with Class
RI	1.5112	1.5339	1.5184	0.0030	-0.1642
Na	10.73	17.38	13.4079	0.8166	0.5030
Mg	0.00	4.49	2.6845	1.4424	-0.7447
Al	0.29	3.50	1.4449	0.4993	0.5988
Si	69.81	75.41	72.6509	0.7745	0.1515
K	0.00	6.21	0.4971	0.6522	-0.0100
Ca	5.43	16.19	8.9570	1.4232	0.0007
Ba	0.00	3.15	0.1750	0.4972	0.5751
Fe	0.00	0.51	0.0570	0.0974	-0.1879

Class Distribution (214 Instances)

Window Glass (163 total)

—Float processed: 87
—Building windows: 70
—Vehicle windows: 17
—Non-float processed: 76
—Building windows: 76
—Vehicle windows: 0

Non-Window Glass (51 total)

—Containers: 13
—Tableware: 9
—Headlamps: 29

Use Case

This dataset is widely used for:

—Machine learning classification tasks
—Forensic science research
—Pattern recognition and data mining studies