Pausanias Analysis

Tri-marked sentence-level mythic vs. historical analysis

Surface, including books 4 and 8, without rhetoric markers

This model drops sentences tagged other and fits a balanced TF-IDF logistic regression to the remaining mythic and historical tags.

Model Performance Metrics

The following metrics are from the logistic regression classifier's performance on the test set:

Class Precision Recall F1-Score Support
Historical 0.714 0.732 0.723 859
Mythic 0.738 0.720 0.729 899
Overall Accuracy 0.726 1758

Confusion Matrix

Predicted
Historical Mythic
Actual Historical 629 230
Mythic 252 647

Counts

7,032mythic/historical sentences
3,598mythic
3,434historical
1,000features

Predictors of Mythic Sentences

Sort by:
Word/Phrase English Coefficient Mythic Count Historical Count p-value q-value
εἶναι to be 3.1026 430 138 7.46e-36 7.46e-33
ἀφικέσθαι to arrive 2.3926 42 4 6.51e-09 4.07e-07
ἔπη words 2.1018 49 2 1.67e-12 2.38e-10
ὕδωρ water 2.0822 57 11 1.35e-08 6.45e-07
αὐτὴν her 1.9285 110 39 1.09e-08 5.44e-07
αὐτήν her 1.8331 33 5 2.86e-06 6.98e-05
θυγατέρα daughter 1.8286 82 9 8.31e-16 4.15e-13
ἀνθρώπων of humans 1.7015 54 25 0.00185 0.0101
γυναῖκα woman; wife 1.6714 59 14 1.19e-07 3.82e-06
ἐστιν is 1.6658 258 116 7.02e-13 1.17e-10
γενέσθαι to become 1.6589 201 83 6.97e-12 8.72e-10
λέγοντεσ speaking 1.5019 31 3 8.81e-07 2.45e-05
καλεῖται is called 1.4949 19 1 4.25e-05 0.000545
βωμὸσ altar 1.4758 19 0 4.08e-06 9.06e-05
θηρίον wild animal 1.4615 17 0 1.61e-05 0.00024
ἐστὶν is 1.4577 140 65 4.61e-07 1.36e-05
καὶ εἶναι and to be 1.4461 42 3 1.12e-09 8.59e-08
ἔστι there is 1.4407 99 36 1.08e-07 3.81e-06
πρῶτον first; first of all 1.4367 60 33 0.00895 0.0335
δʼ but; and 1.4245 109 34 4.95e-10 4.5e-08
θυγατρὶ to/for the daughter 1.4071 18 0 8.1e-06 0.000153
χωρίωι in a place 1.3612 16 1 0.000286 0.00247
ἔστιν there is 1.3422 53 27 0.0061 0.0258
συγγενέσθαι to associate with 1.3392 17 0 1.61e-05 0.00024
εἶναι καὶ also to be 1.3380 78 10 3.97e-14 9.93e-12
αὐτῆι to her 1.3271 56 20 5.45e-05 0.000656
ἔνθα there; where; then 1.3235 75 33 0.000102 0.00112
πεποίηται has been made 1.3174 57 23 0.000235 0.00213
ἀφʼ from 1.3131 28 10 0.00443 0.02
γυναικῶν of women 1.3072 28 10 0.00443 0.02

Predictors of Historical Sentences

Sort by:
Word/Phrase English Coefficient Mythic Count Historical Count p-value q-value
ἐναντία opposite; contrary -2.2999 2 48 4.44e-13 8.88e-11
πόλεμον war -2.2071 12 80 8.47e-15 2.82e-12
πολέμωι in war; by/with war -2.1256 7 53 7.25e-11 8.05e-09
αὖθισ again -1.9907 35 100 1.6e-09 1.14e-07
ἐπίγραμμα epigram; inscription -1.9498 15 34 0.00348 0.0167
εἰκόνα image -1.9258 5 33 7.31e-07 2.09e-05
χρημάτων of money; of goods -1.9101 5 36 1.18e-07 3.82e-06
χρήματα money -1.8333 4 39 9.42e-09 4.96e-07
ἔργου of work -1.7921 4 35 1.11e-07 3.81e-06
πολέμου of war -1.7208 21 73 7.7e-09 4.42e-07
στρατιᾶι to/for the army -1.7080 3 27 3.04e-06 7.13e-05
τούσ the -1.6600 4 28 7.44e-06 0.000146
χαλκοῦν of bronze; brazen -1.6304 8 24 0.00245 0.0128
ἀναστάτουσ expelled -1.5345 1 21 3.7e-06 8.41e-05
πυγμῆι with the fist -1.5344 1 14 0.000396 0.00307
ἄνδρασ men -1.5198 16 53 1.71e-06 4.49e-05
ἐγένετο it happened -1.4985 61 118 3.07e-06 7.13e-05
τούτουσ these -1.4949 14 43 3.77e-05 0.000496
ὅσον as much as -1.4509 17 43 0.000311 0.00264
βουλῆσ of the council; of the will/plan -1.4337 1 18 2.78e-05 0.000381
καὶ περὶ and about -1.4326 9 26 0.00208 0.0113
βασιλέων of kings -1.4288 3 19 0.000375 0.00303
ἐνίκα he/she/it was winning -1.4133 1 30 7.95e-09 4.42e-07
πλέον more -1.4116 18 51 1.99e-05 0.000285
συμμαχίαν alliance -1.3998 1 19 1.42e-05 0.000226
τηνικαῦτα at that very time (just then) -1.3821 18 40 0.00185 0.0101
ἐν μάχηι in battle -1.3802 10 24 0.00996 0.0364
ἀνείλετο he/she/it killed -1.3719 1 19 1.42e-05 0.000226
ἑκατόν one hundred -1.3621 0 16 1.03e-05 0.000183
βαρβάρων of the barbarians -1.3485 7 32 1.58e-05 0.00024