Inkinga eyinhloko ehlobene namakhodi wezibalo ze-inferential ku-Python ukuthi kungase kube nzima ukuqonda nokuhumusha imiphumela. I-Python ulimi olunamandla, kodwa kungase kube nzima ukufunda nokuqonda ikhodi esetshenziselwa izibalo ezingenangqondo. Ukwengeza, kunamaphakheji amaningi ahlukene atholakalayo ezibalo ezingenangqondo kuPython, angenza kube nzima ukukhetha elilungile lokuhlaziya okuthile. Okokugcina, amanye alawa maphakheji angase angabi sesikhathini samanje noma athembeke njengamanye, ngakho-ke kubalulekile ukwenza ucwaningo ngaphambi kokuwasebenzisa.
1. Chi-Square Test of Independence: from scipy.stats import chi2_contingency chi2, p, dof, expected = chi2_contingency(observed) 2. One-Way ANOVA: from scipy import stats F, p = stats.f_oneway(sample1, sample2, sample3) 3. Pearson’s Correlation Coefficient: from scipy.stats import pearsonr corr, _ = pearsonr(x, y)
Umugqa 1: Lo mugqa ungenisa ngaphandle umsebenzi we-chi2_contingency usuka kulabhulali ye-scipy.stats, bese uwusebenzisela ukubala ukuhlola kwesikwele se-chi sokuzimela kudatha ebhekiwe. Imiphumela yalokhu kuhlola igcinwa kokuguquguqukayo i-chi2, p, dof, kanye nelindelwe.
Umugqa wesi-2: Lo mugqa ungenisa umsebenzi we-f_oneway usuka kulabhulali ye-scipy, bese uwusebenzisela ukubala i-ANOVA yendlela eyodwa kumasampuli amathathu (isampula1, isampula2, isampula3). Imiphumela yalokhu kuhlola igcinwe kokuguquguqukayo F kanye ne-p.
Umugqa wesi-3: Lo mugqa ungenisa umsebenzi we-pearsonr usuka kulabhulali ye-scipy.stats, bese uwusebenzisela ukubala i-coefficient yokuhlobanisa ka-Pearson phakathi kokuhluka okubili (x kanye no-y). Imiphumela yalokhu kuhlolwa igcinwa kuma-variables corr kanye ne-_.
Ziyini izibalo ze-inferential
Izibalo ze-inferential ziyigatsha lezibalo elisebenzisa idatha esuka kusampula ukwenza okucatshangwayo noma okuvamile mayelana nenani labantu. Kubandakanya ukwenza iziphetho mayelana nenani labantu ngokusekelwe kudatha eqoqwe kusampula. Ku-Python, izibalo ezingenangqondo zingasetshenziswa ukuze kufinyelelwe esiphethweni nokwenza izibikezelo ngokusebenzisa amasu ahlukahlukene njengokuhlola i-hypothesis, ukuhlaziya ukuhlobana, ukuhlaziywa kokuhlehla, nokuningi. Lawa maqhinga asivumela ukuthi sithole imininingwane ephusile kudatha yethu futhi asisize senze izinqumo ezingcono.
Izinhlobo zezibalo ze-inferential
Ku-Python, kunezinhlobo ezimbalwa zezibalo ezingavamile ezingasetshenziswa ukuhlaziya idatha. Lokhu kufaka phakathi ukuhlolwa kwe-t, i-ANOVA, ukuhlolwa kwe-chi-square, ukuhlolwa kokuhlobana, nokuhlaziywa kokuhlehla. Ukuhlolwa kwe-T kusetshenziselwa ukuqhathanisa izindlela zamaqembu amabili noma ngaphezulu edatha. I-ANOVA isetshenziselwa ukuqhathanisa izindlela zamaqembu amaningi wedatha. Ukuhlolwa kwe-Chi-square kusetshenziselwa ukuhlola ubudlelwano phakathi kokuhluka kwezigaba. Ukuhlolwa kokuhlobana kukala amandla nesiqondiso sobudlelwano bomugqa phakathi kokuhluka okubili. Okokugcina, ukuhlaziywa kokuhlehla kusetshenziselwa ukubikezela ukuguquguquka okuncikile kokuguquguqukayo okukodwa noma ngaphezulu okuzimele.
Uzibhala kanjani izibalo ezingenangqondo
Izibalo ze-inferential ziyigatsha lezibalo elisebenzisa idatha esuka kusampula ukwenza okucatshangwayo mayelana nenani labantu isampula ethathwe kulo. Ku-Python, izibalo ezingenangqondo zingenziwa kusetshenziswa imitapo yolwazi ehlukahlukene njenge-SciPy, i-StatsModels, ne-NumPy.
Ukuze wenze izibalo ze-inferential ku-Python, uzodinga kuqala ukungenisa imitapo yolwazi edingekayo bese usebenzisa imisebenzi efana ne-mean(), median(), mode(), variance(), standard deviation(), t-test(), chi -square test() njll. Isibonelo, uma ubufuna ukubala incazelo yedathasethi enikeziwe, ungasebenzisa umsebenzi we-mean() osuka ku-NumPy:
ngenisa i-numpy njenge-np
idatha = [1,2,3,4]
mean_value = np.mean(data)
phrinta(inani_elisho) # Okukhiphayo: 2.5
