Samples from nearly 800 Bronze Age pottery sherds from Euboea, Eastern Boeotia and Eastern Thessaly were analysed together with 9 raw clays from the same areas. The-analysis was carried out in an attempt to identify areas of pottery manufacture, to discover the origin of specific groups of pottery, to relate pottery to, raw clays and to see how far pottery compositions can be associated with, and predicted by, geology. The work was done on the same lines as earlier studies at the Oxford Laboratory and at the British School at Athens. The main analytical technique used was therefore optical emission spectroscopy. Some 25% of the total number of sherds were also analysed by atomic absorption spectrophotometry so that the results obtained by the two techniques could be compared. The interpretation of the results was facilitated by the use of, computer program packages for cluster and discriminant analysis. Both optical emission and atomic absorption analysis resulted in broadly similar groupings although the absolute concentrations were not directly comparable. The groupings obtained after atomic absorption analysis had the narrower concentration ranges. Nine elements were measured by both techniques but in atomic absorption potassium was added and proved; useful as an additional discriminant. Six composition groups were distinguished from the data. One of them was identified as Euboean, 2 as Boeotian and 3 as coming from different regions of Thessaly. The greatest movement of pottery within these areas was from Euboea to Thessaly. No composition group which originated from outside these regions was identified. Six of the 9 raw clays were associated with the prevailing composition group in the area from which they came. It was not possible to predict trends in pottery composition by examination of the local geology.

Samples from nearly 800 Bronze Age pottery sherds from Euboea, Eastern Boeotia and Eastern Thessaly were analysed together with 9 raw clays from the same areas. The-analysis was carried out in an attempt to identify areas of pottery manufacture, to discover the origin of specific groups of pottery, to relate pottery to, raw clays and to see how far pottery compositions can be associated with, and predicted by, geology. The work was done on the same lines as earlier studies at the Oxford Laboratory and at the British School at Athens. The main analytical technique used was therefore optical emission spectroscopy. Some 25% of the total number of sherds were also analysed by atomic absorption spectrophotometry so that the results obtained by the two techniques could be compared. The interpretation of the results was facilitated by the use of, computer program packages for cluster and discriminant analysis. Both optical emission and atomic absorption analysis resulted in broadly similar groupings although the absolute concentrations were not directly comparable. The groupings obtained after atomic absorption analysis had the narrower concentration ranges. Nine elements were measured by both techniques but in atomic absorption potassium was added and proved; useful as an additional discriminant. Six composition groups were distinguished from the data. One of them was identified as Euboean, 2 as Boeotian and 3 as coming from different regions of Thessaly. The greatest movement of pottery within these areas was from Euboea to Thessaly. No composition group which originated from outside these regions was identified. Six of the 9 raw clays were associated with the prevailing composition group in the area from which they came. It was not possible to predict trends in pottery composition by examination of the local geology.

This paper describes how feature extraction on ancient pottery can be combined with recent developments in artificial intelligence to draw up an automated, but still flexible classification system. These features include for instance several dimensions of the vessel's body, ratios thereof, an abstract representation of the overall shape, the shape of vessel segments and the number and type of attachments such as handles, lugs and feet. While most traditional approaches to classification are based on statistical analysis or the search for fuzzy clusters in high-dimensional spaces, we apply machine learning techniques, such as decision tree algorithms and neural networks. These methods allow for an objective and reproducible classification process. Conclusions about the "typability" of data, the evolution of types and the diagnostic attributes of the types themselves can be drawn as well.

This paper describes how feature extraction on ancient pottery can be combined with recent developments in artificial intelligence to draw up an automated, but still flexible classification system. These features include for instance several dimensions of the vessel's body, ratios thereof, an abstract representation of the overall shape, the shape of vessel segments and the number and type of attachments such as handles, lugs and feet. While most traditional approaches to classification are based on statistical analysis or the search for fuzzy clusters in high-dimensional spaces, we apply machine learning techniques, such as decision tree algorithms and neural networks. These methods allow for an objective and reproducible classification process. Conclusions about the "typability" of data, the evolution of types and the diagnostic attributes of the types themselves can be drawn as well.