Oracle, tell me who I want to be

Hello readers. I want to introduce you to a small study that my friend and I conducted a year and a half ago, but in view of some circumstances, the design has dragged on to this point. The question we wanted to consider is whether it is possible to build such a classifier that would help young people with the choice of a profession.


TL; DR We conducted a survey of three psychological tests: the Big Five, emotional intelligence, a test for the profession. They asked a question by whom they work, whether they like the work. Spent a small EDA. Based on the tests, several classifiers were built and the best was compared with a test for the profession, the answers of which were taken into account with a nuance. Dataset posted in the public domain.



Introduction


, , , . , , - , , , , , . , : , “ ”, , . , -, , , . , , , . ? , :


  • 18 , , , , - . , — , .
  • , , , , , . , , - .
  • , , - . -, , , , , . -, . -, , . . -, ?

, , , . , , . - , , . , , , , . , , "" .


, , : , , , , - .


, , .


, , , , , .



, , . Google., — . , — . :


  • , , ;
  • , , , , ;
  • - .

, . , , , , . , - , .


, : , , , , , Woman.ru . , , , , . , , .


263 . , . — . , " ", . , , , . , , — , .



, , , , . , , , , , , . , , , , , . , . , , , : , . , “ , ?” “ ”, , , , “ ”. “ ?” “ ”. 111 .


, IQ, , , . IQ : — .


.



. , . , — , . . , MMPI, 30 . , , , , , – 60 .


, . , . , , , , , .


  • (Extroversion)- , , — .
  • (Agreeableness)- , , .
  • , (Conscientiousness) — .
  • (Neuroticism) — .
  • (Openness to Expirience) — , , .

50 10 . , , , , , .


EQ-


( ) , , , , . , , , — IQ. Emotional Intelligence, , .


  • (Self-Awareness) — , , .
  • (Self-Management) — .
  • (Motivation) — .
  • (Empathy) — , .
  • (Relationship Management) — , , .

25 , .


, , , .



. . . 5 , . , , . , : , ? , , , . , , , , , , , , . , , .


, , .


  • - — , , , , , , , .
  • - — , , , , , , , , , .
  • - — , , , , , , , , , .
  • - — , , , , , , , , .
  • - — , , , , , , , .

, . , , , . , , IT. IT , , . .


30 6 .



EDA


, . . , , .



, 20-25 , , . , 30 . , , — . . , , , . , , . , ? , , ( — , ) . , .
1 — -
2 — -
3 — -
4 — -.
5 — -
6 —
, ( - ), , . , - .
: .


, , , , . .


. , , . :


  • — , . , . .. , , . .
  • — , . , .
  • — - .

. , , — , , , , , .


, , :


  • , , ( ).
  • , , ( ) .

- . “ , ?”



, , , 20-25 . , , , , , 15 , .
, , — . , , . , . , , , .


, , : 83, , , : 41. , , — 124. , ( , , ) . , , .


, 20-25 25. ,


, 20-25, >25
1122
3232

.. 25 , .


, , 25 22, 32.


: 20 25 , , , . , (), ().


:


  • .
  • , , . , , . , , , — .
  • , , , , , , , , , .
  • . , . , , . ? ?

:



, . , , , . , , . relationship_management , . , Opennes to Expirience. . , - , , , .


:



. (motivation) (self_awareness) (self_managemnt), , : , , . (empathy) , , , , .
, , , 0,26, , 0,36. 826 . , , 0,44, 0,63 . , , , , .
,



, , , , . , . (Neurotism) EQ-, .


, . , ( ) ( ). . , .


EQ-




, , , - . , , . , , .



, RandomForest, , , .


4 . ( 15 ). , , . , . , , . . , CrossValidation 3 .


Linear RegressionRandom Forest
Test results only Acc: 0.60 ± 0.05 Acc: 0.60 ± 0.08
With test only, with polynom Acc: 0.63 ± 0.04 Acc: 0.57 ± 0.05
With questions Acc: 0.55 ± 0.04 Acc :0.60 ± 0.04
With questions with polynom Acc: 0.70 ± 0.01 Acc: 0.70 ± 0.04

, , . , , , , . , , . , , . , 0.3 13 . .


valuename
00.916530Q11 Q92
10.739014Openn Q11
20.611891Neuro Q15
30.418799Q11 Q32
40.124221Neuro Q11
50.019474Q11 Q26
60.006716empathy Q7
70.001681Q7 Q46
8-0.152755sign Q11
9-0.156314empathy Q52
10-0.522895empathy Q27
11-0.614839Q3 Q7
12-0.961098self_management Q11

Q# . , ( , ):


EQ-test Q0 — Q23
big five test Q24 — Q73
prof test Q74 — Q103


Q3: ,
Q7:
Q11:
Q15:
Q26:
Q27:
Q32:
Q46:
Q52:
Q92:


Q92 , .



, . , . , — , .


Accuracy
Profession test0.25 ± 0.021
Classifier0.71 ± 0.055


, , . , , , , . .


, , , , , . , , .. , , . , , .


, , , . , .


, :


  • . , 16-18 , ? 16PF 17 20 — ,
  • , . , , .. . , .
  • , , , , , , .
  • IQ , .

. 260 , . , , , . , , .


, , , . , , issues.


Thanks to my friend, Popov Fedor, a student at the Moscow Aviation Institute, for his help in organizing the test and discussions throughout the study.


PS We apologize to the people who were waiting for the test results, but did not provide them with them. We are really very ashamed.


All Articles