search

excieve / dragnet

Catching the big fish
MIT License
2 stars 1 forks source link

garage_wo_car_flag: add parking places from "other" #50

Open pro100olga opened 4 years ago

pro100olga commented 4 years ago

Now flag is raised if:

Suggestion: besides garages, add also "other" estate objects, which have in their description something about parking place or place for a car. Suggested patterns, based on the analysis of the data are:

(((м[і|е]с[т|ц])|(стоянк)|(схов)|(хранен)|(зберіг)|(парк))(.*)((авто)|(трансп)|(машин)))
(((авто)|(трансп)|(машин))(.*)((м[і|е]с[т|ц])|(стоянк)|(схов)|(хранен)|(зберіг)|(парк)))
(гараж)
(парк[у|о]в)
(парк[і|и]нг)
(парко(.*)м[і|е]с[т|ц])
pro100olga commented 4 years ago

Checked on a partial dataset (88K) ~60 mismatches, all True, but should be False

There is a vehicle in declaration: nacp_00d0b1ec-db68-4b74-b480-94f777d2b1fb nacp_00dab510-cefe-4f5b-bdcf-2c1da9e9292c nacp_9901c610-6c3d-419e-bfba-931b291768df nacp_99042d99-0e63-488a-96d0-f837a3850f00 nacp_9929ee56-3e0a-4a08-9e65-d7b2ad69d4a0

dchaplinsky commented 4 years ago

Well, the question here is should we count in vehicles that aren't owned by declarant or family. Current implementation doesn't count them. Let me know if you'd like to change that

On Thu, Apr 23, 2020 at 9:57 PM Olga Makarova notifications@github.com wrote:

Checked on a partial dataset (88K) ~60 mismatches, all True, but should be False

There is a vehicle in declaration: nacp_00d0b1ec-db68-4b74-b480-94f777d2b1fb nacp_00dab510-cefe-4f5b-bdcf-2c1da9e9292c nacp_9901c610-6c3d-419e-bfba-931b291768df nacp_99042d99-0e63-488a-96d0-f837a3850f00 nacp_9929ee56-3e0a-4a08-9e65-d7b2ad69d4a0

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/excieve/dragnet/issues/50#issuecomment-618592554, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABAA4VGFY43FLGFMTTINHLROCFQBANCNFSM4MPKRLSA .

pro100olga commented 4 years ago

Right, I have checked considering ownership. Now 35 mismatches, all True, but should be False: