Gwaji na F a cikin Nazarin Bambancin
Pendahuluan
A cikin binciken kididdiga, ɗaya daga cikin manyan manufofin shine a fahimci ko akwai manyan bambance-bambance tsakanin ƙungiyoyin bayanai. Gwajin F hanya ce da ake amfani da ita don wannan dalili, musamman a cikin mahallin Nazarin Bambancin (ANOVA). Wannan gwajin yana da mahimmanci a cikin nazarin bayanan gwaji, domin yana ba masu bincike damar tantance amincin sakamakon gwaji muddin sun cika wasu zato na ƙididdiga. A cikin wannan labarin, za mu bincika ra'ayi, aikace-aikace, zato, da fassarar gwajin F a cikin nazarin bambance-bambancen.
Asali na Ma'anar Gwajin F
An sanya wa gwajin F suna ne saboda ƙimarsa tana bin rarrabawar F, wanda shine ci gaba da rarraba yiwuwar da ake amfani da shi wajen nazarin bambancin. Ana amfani da rarrabawar F don kwatanta bambancin tsakanin rukuni da bambancin cikin rukuni, wanda ke taimakawa wajen tantance ko akwai babban bambanci tsakanin ma'aunin rukuni.
Muhimman abubuwan da ke cikin gwajin F sune:
1. Bambancin Cikin-Ƙungiya (Bambanci a Cikin Ƙungiyoyi): Yana auna bambancin bayanai a cikin kowace ƙungiya.
2. Bambancin Tsakanin Ƙungiya (Bambanci tsakanin Ƙungiya): Yana auna matsakaicin bambancin tsakanin ƙungiyoyi.
Idan bambancin da ke tsakanin ƙungiyoyi ya fi bambancin da ke tsakanin ƙungiyoyi girma, akwai yiwuwar bambanci na gaske tsakanin ƙungiyoyin.
Amfani da Gwajin F a cikin ANOVA
ANOVA wata dabara ce ta kididdiga da ake amfani da ita don kwatanta hanyoyin ƙungiyoyi fiye da biyu. Akwai nau'ikan ANOVA daban-daban, gami da ANOVA mai hanya ɗaya, ANOVA mai hanya biyu, da sauran bambance-bambancen. Babban bambanci tsakanin su ya dogara da yanayi da adadin abubuwan da aka yi nazari a kansu. A cikin wannan labarin, za mu mayar da hankali kan ANOVA mai hanya ɗaya a matsayin misali mai sauƙi don kwatanta amfani da gwajin F.
Matakan Bincike tare da ANOVA Hanya Ɗaya
1. Tsarin Hasashe:
– Ba komai bane ($H_0$): Ya bayyana cewa dukkan al'umma suna nufin iri ɗaya ne (babu bambanci tsakanin ƙungiyoyi).
– Madadin hasashe ($H_1$): Yana bayyana cewa akwai aƙalla ma'auni ɗaya na yawan jama'a daban-daban.
2. Lissafa kididdigar F:
– Jimlar Murafu (SST):
\[
SST = \sum_{i=1}^{N}(X_i – \bar{X})^2
\]
Yana auna jimlar bambancin da ke cikin bayanan.
– Jimlar Murafuka Tsakanin Rukuni (SSB):
\[
SSB = \sum_{j=1}^{k} n_j (\bar{X_j} – \bar{X})^2
\]
Yana auna bambancin da ke tsakanin ƙungiyoyi.
– Jimlar Murabba'ai a cikin rukuni (SSW):
\[
SSW = \sum_{j=1}^{k} \sum_{i=1}^{n_j} (X_{ij} – \bar{X_j})^2
\]
Yana auna bambancin da ke cikin kowace ƙungiya.
– Lissafa kididdigar F:
\[
F = \frac{\text{MSB}}{\text{MSW}} = \frac{\text{SSB}/(k-1)}{\text{SSW}/(Nk)}
\]
Inda MSB shine Ma'anar Square Tsakanin Ƙungiyoyi da MSW shine Ma'anar Square Tsakanin Ƙungiyoyi.
3. Muhimmancin Darajar:
Bayan mun ƙididdige ƙimar F, mun kwatanta wannan ƙimar da ƙimar mahimmanci ta rarrabawar F bisa ga matakin mahimmanci (\(\alpha\)) da digiri na 'yanci. Idan ƙimar F da aka ƙididdige ta fi ƙimar mahimmanci, za mu ƙi hasashen da ba shi da ma'ana.
Zato na Gwaji na F
Yana da mahimmanci a tuna cewa amfani da gwajin F ya dogara ne akan zato da yawa na asali. Idan ba a cika waɗannan zato ba, sakamakon gwajin F na iya zama mara inganci. Waɗannan zato sun haɗa da:
1. 'Yanci:
Dole ne lura a kowace ƙungiya ta kasance ba ta da alaƙa da juna.
2. Daidaito:
Bayanai a kowace ƙungiya ya kamata su bi rarrabawar da aka saba. Ana iya gwada zato na al'ada ta amfani da gwaje-gwajen al'ada kamar gwajin Shapiro-Wilk, ko kuma ta hanyar zane-zane ta amfani da zane-zanen QQ.
3. Daidaiton bambancin:
Bambancin da ke cikin kowace ƙungiya dole ne ya zama daidai. Ana iya gwada wannan zato ta amfani da gwajin Levene ko gwajin Bartlett.
Idan ba a cika zato na daidaito ko daidaituwar bambancin ba, akwai canje-canje da za a iya yi (kamar logarithms ko tushen murabba'i) ko wasu gwaje-gwaje marasa ma'auni kamar gwajin Kruskal-Wallis H waɗanda ba sa buƙatar waɗannan zato.
Fassarar Sakamako
Bayan gudanar da ANOVA da kuma samun ƙimar F da ƙimar p, mataki na gaba shine a fassara sakamakon. Ga wasu fassarori masu yiwuwa:
1. Idan darajar p <\(\alpha\): An ƙi hasashen null, wanda ke nuna cewa akwai babban bambanci tsakanin ma'anar rukuni. 2. Idan darajar p > \(\alpha\): Babu isassun shaidu don ƙin yarda da ma'anar null, wanda ke nuna cewa babu wani bambanci mai mahimmanci tsakanin ma'anar rukuni.
Ko da an ƙi yarda da hasashen da ba shi da ma'ana, ANOVA ba ta nuna waɗanne ƙungiyoyi suka bambanta ba. Wannan yana buƙatar gwaje-gwajen bayan-hoc kamar Tukey's HSD (Gaskiya Mai Muhimmanci Difference), gyaran Bonferroni, ko gwajin Sidak, wanda zai iya taimakawa wajen gano waɗanne ƙungiyoyi ne suka bambanta sosai.
Samfurin akwati
Bari mu yi la'akari da wannan misali mai sauƙi:
Mai bincike yana son tantance ko akwai babban bambanci a cikin ingancin nau'ikan taki guda uku akan ci gaban shuka. Mai binciken yana auna tsayin shuka (a cm) bayan wata guda ga ƙungiyoyi uku na shuka ta amfani da takin zamani na A, B, da C.
Bayanan Hasashe:
| Takin A | Takin B | Takin C |
|———|———|———|
| 20 | 18 | 22 |
| 21 | 17 | 23 |
| 19 | 16 | 24 |
Matakan Bincike:
1. Tsarin Hasashe:
– $H_0$: Matsakaicin tsayin shuka ga dukkan takin zamani iri ɗaya ne.
– $H_1$: Aƙalla matsakaicin tsayin shuka ɗaya ya bambanta.
2. Lissafa kididdigar F:
– Lissafa SST, SSB, da SSW sannan a ci gaba da lissafin F.
3. Kwatanta da Muhimmancin Daraja:
– Ta amfani da teburin rarrabawa na F da digirin 'yanci, muna tantance ko ƙimar F da aka ƙididdige tana da mahimmanci.
Ƙarshe:
Idan ƙimar F ta nuna cewa akwai babban bambanci, mai binciken zai iya ci gaba da gwaje-gwajen bayan-hoc don tantance waɗanne ƙungiyoyi ne suka bambanta.
Kammalawa
Gwajin F-gwajin a fannin nazarin bambance-bambance kayan aiki ne mai matuƙar amfani don tantance ko akwai manyan bambance-bambance tsakanin ƙungiyoyi. Ta hanyar cimma hasashen ƙididdiga, wannan gwajin zai iya samar da fahimta mai ƙarfi game da bayanan da aka bincika. A aikace-aikace, wannan gwajin yana da matuƙar amfani a fannoni daban-daban na bincike kamar ilmin halitta, nazarin zamantakewa, tattalin arziki, da sauransu. Sanin lokacin da kuma yadda ake amfani da gwajin F-gwajin, da kuma fahimtar zato da fassarorinsa, zai inganta ingancin nazarin ƙididdiga kuma ya samar da tushe mai ƙarfi don yanke shawara bisa ga bayanai.
