Properties of Data Flashcards

1
Q

Was ist der Unterschied zwischen experimental data und observational data?

A

experimental data: entsteht unter kontrollierten Bedingungen
observational data: unkontrollierte Datenansammlung

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Wie bestimmt man den geometrischen Durchschnitt?

A
good for ratio data type
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Im Bezug auf Lagewerte, welche Eigenschaft hat eine symmetrische Verteilung?

A

Median = Mean = Modus

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Wie berechnet man die Varianz?

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Wie berechnet man die Standardabweichung?

A

sqrt(Varianz^2)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Was ist der L1-Loss?

A

minimized by the median m

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Was ist der L2-Loss?

A

minimized by the mean μ

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Was besagt das Central Limit Theorem?

A

the sum of many random variables converges to a Gaussian

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Wie viel Prozent beinhaltet der Bereich unter der Funktion der Normalverteilung von bis ?

A

~68%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Wie viel Prozent beinhaltet der Beich unter der Funktion der Normalverteilung von -2σ bis +2σ?

A

~95%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Wie viel Prozent beinhaltet der Bereich unter der Funktion der Normalverteilung von -3σ bis +3σ?

A

~99.7%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Beschreib Boxplots

A
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Beschreib die Curse of Dimensionality

A
  1. we need exponentially more data for constant density,
  2. a hypercube of larger edge length covers same subspace,
  3. distance between points increases,
  4. distance to an edge decreases,
  5. every point becomes an outlier.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Wie löst man den Curse of Dimensionality?

A

If high-dimensional, we need more data for density estimation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly