Matanho Ekupararira: Kunzwisisa Kuchinja-chinja muData
Mukuongorora nhamba nedata, kunzwisisa kugoverwa nekusiyana kwedata kwakakosha pakuita fungidziro dzakarurama uye dzakakodzera. Imwe pfungwa huru inoshandiswa kutsanangura kusiyana kwedata "chiyero chekupararira." Chinyorwa chino chichakurukura zviyero zvakasiyana-siyana zvekupararira, kuti nei zvichikosha, maitiro ekuzviverenga, uye dudziro yazvo mukuongorora data.
Chii chinonzi Kuyera Kupararira?
Zviyero zvekupararira zviyero zvinoshandiswa kutsanangura kuti data riri museti rinopararira sei kana kupararira kubva pahuwandu hwepakati. Hunhu uhwu hwepakati hunowanzo kuyerwa uchishandisa zviyero zvepakati zvakaita sepakati kana pakati. Zviyero zvekupararira zvinopa ruzivo rwehuwandu, kusiyana, uye kuenderana kwedata.
Sei Kupararira Kwakakosha?
1. Kunzwisisa Kuchinja-chinja:
Kuchinja-chinja chikamu chakakosha chedata chero ripi zvaro. Nekunzwisisa kuti data rinosiyana sei, tinogona kunzwisisa mashandiro ari pasi pedata iroro.
2. Ziva Zvisingakoshi:
Kugoverwa kwedata kunogona kubatsira mukuona zvinhu zvisingawanikwe (zvakanyanya kukosha zviri kure nedzimwe data), izvo zvingave zvakakosha pakuongorora kwakawedzerwa kana kuti zvingave data rezvikanganiso.
3. Kuenzanisa kweDataset:
Kuyera kwekupararira kunobvumira kuenzanisa pakati pemaseti maviri edata kana kupfuura. Semuenzaniso, maseti maviri edata anogona kunge aine avhareji imwe chete asi akasiyana kana kuti kupararira.
4. Nhamba dzekufungidzira:
Nzira dzakawanda dzekufungidzira dzinoda kunzwisisa kwakanaka kwekugoverwa kwedata kuti pave nemhedziso dzakakosha uye dzinoshanda.
Mhando dzeSaizi yeSpread
Kune nzira dzakasiyana-siyana dzekuyera kupararira kwedata (dispersion tests) dzinowanzoshandiswa mukuongorora data rehuwandu hwedata:
1. Nzvimbo
Range ndiyo nzira iri nyore yekuyera kupararira uye inoverengerwa semusiyano uripo pakati pehuwandu hwepamusoro nehudiki mu data set.
\[ \text{Range} = \text{Kukosha kukuru} – \text{Kukosha kushoma} \]
Kunyangwe zviri nyore kuverenga, huwandu hwacho hunongotarisa mapoinzi maviri edata uye hahuratidzi kugoverwa kwedata pakati pemakoshesi mashoma neakanyanya.
2. Interquartile Range (IQR)
IQR inzira yakasimba yekuyera kupararira kupfuura range nekuti haikanganisike ne outliers. Inoverenga medium range yedata nekubvisa 25th percentile (Q1) kubva pa75th percentile (Q3).
\[ \mashoko{IQR} = Mubvunzo 3 – Mubvunzo 1 \]
Nekutarisa paavhareji, IQR inopa mufananidzo uri nani wekugoverwa kwedata riri pasi.
3. Kusiyana
Kusiyana kunoyera kuti kukosha kwega kwega mu data set kuri kure zvakadii kubva paavhareji. Kunoverengerwa nekupfupisa masikweya emusiyano wemutengo wega wega kubva paavhareji, wobva wagovaniswa nenhamba yezvinhu zvedata (yevanhu) kana nhamba yezvinhu kubvisa chimwe (yemuenzaniso).
Kune vanhu (\(\sigma^2\)):
\[ \sigma^2 = \frac{\sum (X_i – \mu)^2}{N} \]
Semuenzaniso (\(s^2\)):
\[ s^2 = \frac{\sum (X_i – \overline{X})^2}{n-1} \]
Kusiyana-siyana kunopa pfungwa yekuenderana kwedata; zvisinei, nekuti kusiyana kunoshandisa mayuniti akaenzana, zvinogona kuoma kududzira zvakananga.
4. Kutsauka Kwakajairwa
Kutsauka kwakajairwa ndiko mudzi wepakati wemusiyano uye kuri muzvikamu zvakafanana nedata rekutanga, zvichiita kuti zvive nyore kududzira.
Kune huwandu hwevanhu (\(\sigma\)):
\[ \sigma = \sqrt{\sigma^2} = \sqrt{\frac{\sum (X_i – \mu)^2}{N}} \]
Semuenzaniso (\(s\)):
\[ s = \sqrt{s^2} = \sqrt{\frac{\sum (X_i – \overline{X})^2}{n-1}} \]
Kutsauka kwakajairwa ndeimwe yenzira dzinonyanya kushandiswa dzekuyera kupararira nekuti zviri nyore kududzira uye zvinowanzoshandiswa mukuongorora kwakasiyana-siyana kwenhamba.
5. Kuchinja kweKoefficient (CV)
CV chiyero chekupararira kwehuwandu chinoratidzwa sechiyero chekutsauka kwakajairwa kune avhareji uye chinowanzo ratidzwa sechikamu.
\[ \text{CV} = \frac{s}{\overline{X}} \times 100\% \]
CV inobatsira zvikuru pakuenzanisa kusiyana pakati pemaseti edata nenzira dzakasiyana.
Maitiro Ekuverenga Uye Kududzira
Muenzaniso weKuverenga
Ngatiratidzei nemuenzaniso wedata unotevera:
\[ \{15, 20, 25, 35, 45, 55, 65, 75, 85, 95\} \]
1. Nzvimbo:
\[ \text{Range} = 95 – 15 = 80 \]
2. Interquartile Range (IQR):
Mushure mekuronga data, tinogona kuwana makota Q1 neQ3. Muchiitiko ichi, Q1 i25 uye Q3 i75.
\[ \mashoko{IQR} = 75 – 25 = 50 \]
3. Kusiyana uye Kutsauka Kwakajairika:
Avhareji (\(\overline{X}\)) yedata i51.5. Zvadaro tinoverenga variance uye standard deviation.
\[ \text{Variance (s^2)} = \frac{1}{n-1} \sum (X_i – \overline{X})^2 = 816.11 \]
\[ \text{Standard Deviation (s)} = \sqrt{816.11} = 28.57 \]
4. Kuchinja kweKoefficient (CV):
\[ \text{CV} = \frac{28.57}{51.5} \times 100\% \approx 55.48\% \]
Kubva pano, tinogona kududzira kuti standard deviation i28.57, nepo CV ichiratidza kuti standard deviation inenge 55.48% yeavhareji yedata rekutanga.
Mhedziso
Kuyera kwekupararira zvinhu zvakakosha pakuongorora data rehuwandu nekuti zvinopa ruzivo rwekusiyana uye kupararira kwedata rakatenderedza kukosha kwepakati. Kuyera kwekupararira kunowanzo shandiswa kunosanganisira range, interquartile range, variance, standard deviation, uye coefficient of variation. Chimwe nechimwe chezviyero izvi chine mashandisirwo azvo uye chinogona kupa ruzivo rwakakosha zvichienderana nemamiriro edata nechinangwa chekuongorora. Nekunzwisisa nekushandisa zviyero zvekupararira zvakakodzera, tinogona kuita sarudzo dzakanyatsojeka uye dzakarurama muminda yakasiyana-siyana yekutsvagisa uye mashandisirwo esainzi yedata.
