Skip to main content

Table 2 Prediction of plant species in each blind clay pipe (BCP)

From: Utilization of a natural language processing-based approach to determine the composition of artifact residues

Blind clay pipe (BCP)

Standard method (number of mass spectral features shared with the experimental clay pipe)

Key

New method (similarity score)

Second method (similarity score)

Third method (similarity score)

BCP1

Auv (3)

Nta

Nta (0.0370)

Nta (0.0301)

Nta (0.0539)

BCP2

Nat (6) and Nqu (6)

Nqu

Nqu (0.1102)

Nqu (0.1066)

Nqu (0.1400)

BCP3

Nob (10)

Nob

Nob (0.1132)

Nob (0.1107)

Ngl (0.1499)

BCP4

Nat (4)

Alu

Nta (0.0928)

Nta (0.0875)

Nta (0.1336)

BCP5

Ngl (6)

Ngl

Ngl (0.0874)

Ngl (0.0799)

Ngl (0.1103)

BCP6

Lin (10)

Lin

Ngl (0.0844)

Ngl (0.0749)

Ngl (0.1452)

BCP7

Auv (16)

Auv

Auv (0.0792)

Auv (0.0800)

Cse (0.1128)

BCP8

Auv (10) and Cse (5)

Auv and Nta mixture

Cse (0.0853) and Auv (0.0662)

Cse (0.0778) and Auv (0.0643)

Cse (0.1336) and Auv (0.0951)

  1. Clay pipes experimentally smoked with a single, known plant species or AmSp were compared individually to each blind clay pipe (BCP) to determine which plant species had been smoked in a BCP. The expected results are the plant species listed under the column labeled “Key”. The second method is where tf was replaced with 1 + log(tf), and the third method is where idf = 1. Alu, Artemisia ludoviciana; Auv, Arctostaphylos uva-ursi; Cse, Cornus sericea; Lin, Lobelia inflata; Nat, Nicotiana attenuata; Ngl, Nicotiana glauca; Nob, Nicotiana obtusifolia; Nqu, Nicotiana quadrivalvis; Nta, Nicotiana tabacum; and AmSp, American Spirit