Daalụ maka ileta Nature.com. Ụdị ihe nchọgharị ị na-eji nwere nkwado dị nta maka CSS. Maka ahụmịhe kachasị mma, anyị na-akwado ka ị jiri ihe nchọgharị emelitere (ma ọ bụ gbanyụọ ọnọdụ ndakọrịta na Internet Explorer) .Ka ọ dị ugbu a, iji hụ na nkwado na-aga n'ihu, anyị ga-egosipụta saịtị na-enweghị ụdị na Javascript.
Mmetọ ala bụ nnukwu nsogbu nke ihe omume mmadụ na-eme. Nkesa mbara igwe nke ihe nwere ike ime ihe na-egbu egbu (PTE) dịgasị iche iche n'ọtụtụ obodo na mpaghara ime obodo. Ya mere, ọ na-esiri ike ịkọ amụma ọdịnaya nke PTE na ala ndị dị otú ahụ. Ngụkọta nke 115 samples nwetara site na Frydek Mistek na Czech Republic. Calcium (Calcium) (Calcium) (Calcium) (Calcium) (Calcium) na potassium (Ca), magnesium (Ca), magnesium (Calcium) (Calcium) (Calcium) (Calcium) (Calcium) (Calcium) (Calcium) (Calcium) (Calcium) na potassium (Ca). Inductively jikọtara plasma emission spectrometry. Nzaghachi nzaghachi bụ Ni na ndị na-ebu amụma bụ Ca, Mg, na K. Mgbanwe njikọ dị n'etiti mgbanwe nzaghachi na mgbanwe amụma na-egosi mmekọrịta na-eju afọ n'etiti ihe ndị ahụ. Nsonaazụ amụma gosiri na Nkwado Vector Machine Regression (SVMR) rụrụ nke ọma, ọ bụ ezie na atụmatụ ya mere atụmatụ mgbọrọgwụ pụtara njehie square (RMSE) na njehie 235. (166.946 mg / kg) dị elu karịa ụzọ ndị ọzọ a na-etinye. Ngwakọta ngwakọta maka Empirical Bayesian Kriging-Multiple Linear Regression (EBK-MLR) na-eme nke ọma, dị ka ihe àmà na-egosi na ọnụọgụ nke mkpebi siri ike na-erughị 0.1. The Empirical Bayesian Kriging-Support Vector Machine Regression (EBK-MLR) bụ ihe nlereanya kachasị mma. (95.479 mg / kg) na MAE (77.368 mg / kg) ụkpụrụ na ọnụ ọgụgụ dị elu nke mkpebi siri ike (R2 = 0.637) EBK-SVMR usoro mmepụta a na-ahụ anya site na iji map na-ahazi onwe. Ona. Nsonaazụ gosiri na ijikọta EBK na SVMR bụ usoro dị irè maka ịkọ ọkwa Ni n'ime obodo mepere emepe na ala ime obodo.
A na-ewere Nickel (Ni) dị ka micronutrient maka osisi n'ihi na ọ na-enye aka na nhazi nke ikuku nitrogen (N) na urea metabolism, nke abụọ a chọrọ maka mkpụrụ germination.Na mgbakwunye na ntinye aka ya na mkpụrụ germination, Ni nwere ike ime ihe dị ka onye na-eme ihe na-emepụta ihe na nje bacteria ma kwalite mmepe osisi. Enweghị nickel n'ime ala na-enye ohere ka osisi ahụ nwee ike ịmịnye ya, na-eme ka ihe atụ nke akwụkwọ ndụ akwụkwọ ndụ chlorosis na-akpata. nke nickel dabeere na fatịlaịza na-ebuli nitrogen fixation2.Na-aga n'ihu ngwa nke nickel dabeere fatịlaịza na-eme ka ala na-amụba ikike nke mkpo ọkụ na-edozi nitrogen na ala nọgidere na-abawanye nickel ịta n'ime ala. Ọ bụ ezie na nickel bụ a micronutrient maka osisi, ya oké oriri na ala nwere ike ime ihe ọjọọ karịa mma. dị ka ihe dị mkpa na-edozi ahụ maka uto osisi1. Dị ka Liu3 si kwuo, a chọpụtala Ni na ọ bụ ihe dị mkpa nke 17 chọrọ maka mmepe na mmepe osisi.Na mgbakwunye na ọrụ nickel na mmepe osisi na mmepe, ụmụ mmadụ chọrọ ya maka ngwa dị iche iche.Electroplating, mmepụta nke nickel dabeere alloys, na mmepụta nke mgbanye ngwaọrụ na spark plugs na mgbakwunye na ụlọ ọrụ mmepụta ihe niile na-achọ ka ụlọ ọrụ mmepụta ihe dị iche iche. nickel dabeere alloys na electroplated isiokwu e ọtụtụ-eji na kichinware, ballroom ngwa, nri ụlọ ọrụ onunu, eletriki, waya na cable, jet turbines, ịwa implants, textiles, na shipbuilding5.Ni-ọgaranya etoju na ala (ie, elu Ona) e ekewet ma anthropogenic na eke isi mmalite, ma isi, Ni bụ a eke4 isi mmalite nke nickel 6. mgbawa, ahịhịa, ọkụ ọhịa, na usoro ala; Otú ọ dị, isi mmalite anthropogenic gụnyere batrị nickel / cadmium na ụlọ ọrụ nchara, electroplating, arc welding, diesel na mmanụ mmanụ, na ikuku ikuku sitere na combustion coal na mkpofu na sludge incineration Nickel accumulation7,8. Dị ka Freedman na Hutchinson9 na Manyiwa et al. 10, isi mmalite nke mmetọ n'elu ala na nso nso na gburugburu ebe obibi bụ tumadi nickel-ọla kọpa dabeere smelters na mines. The top ala gburugburu Sudbury nickel-ọla refinery na Canada nwere kasị elu ọkwa nke nickel mmetọ na 26,000 mg / kg11. N'ụzọ dị iche, mmetọ si Russia ka nickel mmepụta na Norwegian mmepụta nickel. ala11. Dị ka Alms et al. 12, ego nke HNO3-extractable nickel na mpaghara n'elu ala ubi (nickel mmepụta na Russia) sitere na 6.25 ruo 136.88 mg / kg, kwekọrọ na nke 30.43 mg / kg na a baseline ịta nke 25 mg / kg. Dị ka kabata 11 nke fatịlaịza site n'ime obodo nke fatịlaịza nke fatịlaịza n'ime ime obodo ma ọ bụ site na fatịlaịza 11 nke phosphorus. n'oge oge ihe ọkụkụ na-aga n'ihu nwere ike ịmịnye ma ọ bụ mebie ala. Mmetụta nke nickel nwere ike ịkpata ọrịa cancer site na mutagenesis, mmebi nke chromosomal, ọgbọ Z-DNA, igbochi DNA excision mmezi, ma ọ bụ usoro epigenetic.
Ntụle mmetọ nke ala na-aga n'ihu n'oge na-adịbeghị anya n'ihi ọtụtụ nsogbu ahụike metụtara ahụike sitere na mmekọrịta ala-osisi, ala na mmekọrịta ndụ nke ala, mmebi gburugburu ebe obibi, na ntule mmetụta gburugburu ebe obibi. Ka ọ dị ugbu a, amụma gbasara ohere nke ihe ndị nwere ike igbu egbu (PTE) dị ka Ni n'ime ala na-arụsi ọrụ ike ma na-ewe oge site n'iji usoro omenala dijitalụ nke ọma ugbu a. amụma ala nkewa (PSM) .Dịka Minasny na McBratney16, amụma ala nkewa (DSM) egosila na ọ bụ a ma ama subdiscipline nke ala sayensị.Lagacherie na McBratney, 2006 kọwaa DSM dị ka "eke na njuputa nke gbasara mbara ala ozi usoro site na iji na ọnọdụ na laabu observational usoro na usoro na-abụghị nke ala na-ahụ maka ụlọ nyocha na usoro na-abụghị nke ala na usoro". al. 17 na-akọwapụta na DSM ma ọ bụ PSM nke oge a bụ usoro kachasị dị irè maka ịkọ ma ọ bụ maapụ nkesa mbara igwe nke PTE, ụdị ala na ihe onwunwe ala.Geostatistics and Machine Learning Algorithms (MLA) bụ usoro nhazi nke DSM nke na-emepụta maapụ digitized site n'enyemaka nke kọmputa na-eji data dị ịrịba ama na nke ntakiri.
Deutsch18 na Olea19 na-akọwa geostatistics dị ka "nchịkọta usoro ọnụọgụgụ nke na-ekwu maka nnochite anya njirimara oghere, na-ejikarị ụdị stochastic eme ihe, dị ka otu nyocha usoro oge si akọwa data nwa oge." N'ụzọ bụ isi, geostatistics gụnyere nyocha nke variograms, nke na-enye ohere Quantify na kọwapụta ndabere nke ụkpụrụ oghere site na dataset ọ bụla20.Gumiaux et al. 20 na-egosikwa na nyocha nke variograms na geostatistics dabere na ụkpụrụ atọ, gụnyere (a) ịgbakọ ọnụ ọgụgụ nke njikọ data, (b) ịchọpụta na ịgbakọ anisotropy na disparity dataset na (c) na mgbakwunye na iburu n'uche njehie dị n'ime nke data nha nke kewapụrụ na mmetụta mpaghara, a na-eji ọtụtụ mmetụta eme ihe na mpaghara a. geostatistics, gụnyere kriging izugbe, ngalaba-kriging, kriging nkịtị, empirical Bayesian kriging, usoro kriging dị mfe na usoro mmekọrịta ndị ọzọ ama ama nke ọma iji maapụ ma ọ bụ ịkọ PTE, njirimara ala na ụdị ala.
Machine Learning Algorithms (MLA) bụ a dịtụ ọhụrụ Usoro na-ewe ndị ibu na-abụghị linear data klas, fueled site algọridim bụ isi eji maka data Ngwuputa, na-achọpụta ụkpụrụ na data, na ugboro ugboro etinyere na nhazi ọkwa na sayensị ubi dị ka ala sayensị na nloghachi tasks.Ọtụtụ nnyocha akwụkwọ na-adabere na MLA ụdị ka amụma PTE na ala, dị ka Tan et al. 22 (oke ọhịa na-enweghị usoro maka ntule igwe dị arọ na ala ugbo), Sakizadeh et al. 23 (ihe nlere anya site na iji igwe nkwado vector na netwọk akwara artificial) mmetọ ala .Na mgbakwunye, Vega et al. 24 (CART maka ịmegharị njigide ọla dị arọ na mgbasa ozi na ala) Sun et al. 25 (ngwa nke cubist bụ nkesa nke Cd na ala) na algọridim ndị ọzọ dị ka onye agbata obi kacha nso, nkwụghachi azụ agbagoro agbagoro, na nkwụghachi azụ Osisi tinyekwara MLA iji buo PTE na ala.
Ngwa nke DSM algọridim na amụma ma ọ bụ nkewa na-eche ọtụtụ nsogbu ihu. Ọtụtụ ndị na-ede akwụkwọ kwenyere na MLA dị elu karịa geostatistics na nke ọzọ. Ọ bụ ezie na otu dị mma karịa nke ọzọ, nchikota nke abụọ ahụ na-eme ka ọkwa nke ziri ezi nke eserese ma ọ bụ amụma dị na DSM15.Woodcock na Gopal26 Finke27; Pontius na Cheuk28 na Grunwald29 na-ekwu banyere adịghị ike na ụfọdụ njehie na amụma ala nkewa. Ndị ọkà mmụta sayensị ala agbalịwo usoro dị iche iche iji kwalite ịdị irè, izi ezi, na amụma amụma nke DSM maping na amụma. Nchikota nke ejighị n'aka na nkwenye bụ otu n'ime ọtụtụ akụkụ dị iche iche agbakwunyere na DSM iji mee ka arụmọrụ dị mma na belata. 15 na-akọwapụta na omume nkwado na ejighị n'aka ewepụtara site na imepụta map na amụma kwesịrị ịkwado onwe ya iji melite ogo map. Oke nke DSM bụ n'ihi ọdịdị ala gbasasịrị agbasasị na mpaghara, nke gụnyere akụkụ nke ejighị n'aka; Otú ọ dị, enweghị nke doro anya na DSM nwere ike ibili site na ọtụtụ isi mmalite nke njehie, ya bụ covariate njehie, nlereanya njehie, ebe njehie, na analytical Error 31.Modelling na-ezighị ezi ebutere na MLA na geostatistical Filiks na-ejikọta ya na enweghị nghọta, n'ikpeazụ na-eduga oversimplification nke ezigbo process32. N'agbanyeghị ọdịdị nke ihe nlereanya nke nwere ike ime ihe nlereanya, na-enweghị ike ime ihe nlereanya, na-enweghị ike ịme ihe nlereanya nke ụdị ihe nlereanya ahụ. amụma, ma ọ bụ interpolation33. N'oge na-adịbeghị anya, usoro DSM ọhụrụ apụtala nke na-akwalite njikọ nke geostatistics na MLA na nkewa na ịkọ amụma. Ọtụtụ ndị ọkà mmụta sayensị na ndị na-ede akwụkwọ, dị ka Sergeev et al. 34; Subbotina et al. 35; Tarasov et al. 36 na Tarasov et al. 37 ejirila ezigbo geostatistics na mmụta igwe iji mepụta ụdị ngwakọ na-eme ka arụmọrụ amụma na eserese dị mma. àgwà.Ụfọdụ n'ime ụdị ngwakọ ndị a ma ọ bụ ngwakọta algọridim bụ Artificial Neural Network Kriging (ANN-RK), Multilayer Perceptron Residual Kriging (MLP-RK), Generalized Regression Neural Network Residual Kriging (GR- NNRK) 36, Artificial Neural Network Kriging-Multilayer Perceptron (ANN-K-7 na Coussi-MLP) 36. Ntughari38.
Dị ka Sergeev et al., ijikọta dị iche iche modeling usoro nwere ike ikpochapụ ntụpọ na-amụba arụmọrụ nke dapụtara ngwakọ nlereanya kama ịzụlite ya otu model.Na nke a gburugburu, a ọhụrụ akwụkwọ na-arụ ụka na ọ dị mkpa na-etinye a jikọtara algọridim nke geostatistics na MLA na-emepụta ezigbo ngwakọ ụdị ka amụma Ni enrichment n'ime obodo na peri-urban nke obodo Bayesi (Kesisi) Nke a ga-eme ka ọmụmụ ihe na-eme n'ime obodo na mpaghara nke KESE. dị ka isi nlereanya na mix ya na Nkwado Vector Machine (SVM) na Multiple Linear Regression (MLR) ụdị.Hybridization nke EBK na ihe ọ bụla MLA bụ amaghị.The multiple agwakọta ụdị hụrụ bụ nchikota nke nkịtị, residual, regression kriging, na MLA.EBK bụ a geostatistical interpolation usoro na utilizes a spatially interpolation usoro na-n'ógbè a na-abụghị nke ubi usoro na-na-stochastic usoro na-stochastic ubi. akọwapụtara mpaghara mpaghara n'ọhịa, na-enye ohere maka mgbanwe dị iche iche39.EBK ejirila ya n'ọtụtụ ọmụmụ, gụnyere nyocha nkesa carbon organic na ala ugbo40, na-enyocha mmetọ ala41 na eserese ala42.
N'aka nke ọzọ, Self-Organizing Graph (SeOM) bụ mmụta mmụta nke etinyere n'akwụkwọ dị iche iche dị ka Li et al. 43, Wang et al. 44, Hossain Bhuiyan et al. 45 na Kebonye et al.46 Kpebie àgwà gbasara mbara igwe na nhazi nke ihe.Wang et al. 44 na-akọwapụta na SeOM bụ usoro mmụta dị ike nke a maara maka ikike ijikọta na icheta nsogbu ndị na-abụghị nke linear. N'adịghị ka usoro njirimara ndị ọzọ dị ka nyocha nke isi akụrụngwa, nchịkọta na-enweghị isi, nchịkọta nhazi, na ime mkpebi dị iche iche, SeOM ka mma n'ịhazi na ịchọpụta ụkpụrụ PTE. Dị ka Wang et al. 44, SeOM nwere ike ikpokọta nkesa nke neurons ndị metụtara ya wee nye nhụta data dị elu.
Akwụkwọ a aims n'ịwa a siri ike nkewa nlereanya na ezigbo ziri ezi maka ịkọ ọdịnaya nickel n'ime obodo na mpụta obodo ala. Anyị na-eche na a pụrụ ịdabere na nke ngwakọta nlereanya tumadi dabere na mmetụta nke ndị ọzọ ụdị mmasị na isi model. Anyị na-ekweta ihe ịma aka chere DSM ihu, na mgbe nsogbu ndị a na-egbo n'ọtụtụ ihu, Nchikota ọganihu na geostatistics na-egosi na MLA. Ya mere, anyị ga-anwa ịza ajụjụ nyocha nke nwere ike inye ụdị ụdị agwakọta. Otú ọ dị, olee otú ihe nlereanya ahụ si buru amụma ihe mgbaru ọsọ? Ọzọkwa, gịnị bụ ọkwa nyocha nke arụmọrụ dabere na nkwenye na nyocha ziri ezi? Ya mere, ihe mgbaru ọsọ dị iche iche nke ọmụmụ ihe a bụ (a) ịmepụta ngwakọta ngwakọta maka SVMR ma ọ bụ MLR site na iji EBK dị ka ihe nlereanya nke isi, (b) na-atụnyere ihe nlereanya kachasị mma na-atụnyere ihe nlereanya na-emepụta ihe n'ịtụle ihe atụ nke ngwakọta nke Nii. ala obodo mepere emepe ma ọ bụ obodo dịpụrụ adịpụ , na (d) ntinye nke SeOM iji mepụta maapụ dị elu nke mgbanwe nickel spatial.
A na-eme ọmụmụ ihe a na Czech Republic, na mpaghara Stridek nke Frydek (nke bụ akụkụ nke mpaghara na-eme ihe dị ka 49 '0' n na 18 Celsius 20 ' 0 'e, na elu dị n'etiti 225 na 327 m; Otú ọ dị, usoro nhazi nke Koppen maka ọnọdụ ihu igwe nke mpaghara ahụ bụ Cfb = ihu igwe na-ekpo ọkụ nke oké osimiri, E nwere ọtụtụ mmiri ozuzo ọbụna na ọnwa akọrọ. Okpomọkụ na-adịgasị iche iche n'ime afọ n'etiti -5 °C na 24 Celsius C, ọ na-adịkarịghị ịda n'okpuru −14 °C ma ọ bụ n'elu 30 Celsius C, ebe nkezi nke 7 na preci 6.5. mpaghara nke dum ebe bụ 1,208 square kilomita, na 39.38% nke akọ ala na 49.36% nke oke ohia mkpuchi. N'aka nke ọzọ, ebe a na-eji na ọmụmụ ihe bụ banyere 889.8 square kilomita. In na gburugburu Ostrava, ígwè ọrụ na ígwè ọrụ na-arụ ọrụ nke ukwuu. Metal igwe igwe, ígwè na-eji ígwè na-eguzogide ígwè na igwe igwe anaghị agba nchara (eg. corrosion) na alloy steels (nickel enwekwu ike nke alloy mgbe ịnọgide na-enwe ezi ductility na siri ike), na kpụ ọkụ n'ọnụ agriculture dị ka phosphate fatịlaịza ngwa na anụ ụlọ mmepụta bụ nnyocha nwere ike isi mmalite nke nickel na mpaghara (eg, na-agbakwụnye nickel na atụrụ na-abawanye ibu udu na ụmụ atụrụ na ala-azụ ehi) .Ndị ọzọ ulo oru eji electroplating gụnyere nickel. electroless nickel plating processes. Ala Njirimara na-adị mfe iche site na agba agba, Ọdịdị, na carbonate content.The ala udidi bụ ọkara na ezi, ewepụtara site na nne na nna material.They bụ colluvial, alluvial ma ọ bụ aeolian na nature.Some ala ebe na-egosi mottled na elu na ala ala, mgbe na ihe na bleaching. Otú ọ dị, ala cambisols na ndị kasị nkịtị n'ógbè8. sitere na 455.1 ruo 493.5 m, cambisols na-achị Czech Republic49.
Maapụ mpaghara ọmụmụ [E ji ArcGIS Desktop mepụta maapụ mpaghara ọmụmụ (ESRI, Inc, ụdị 10.7, URL: https://desktop.arcgis.com).]
A na-enweta ngụkọta nke 115 n'elu ala site na obodo ukwu na obodo nta dị na mpaghara Frydek Mistek. Ụdị ihe atụ a na-eji eme ihe bụ grid mgbe niile na ihe atụ nke ala dị n'ebe dịpụrụ adịpụ 2 × 2 km, a na-atụkwa elu ala n'ime omimi nke 0 ruo 20 cm site na iji ngwaọrụ aka (Leica Zeno 5 GPS, nke a na-agbanye n'ụdị GPS, nke a na-agbanye n'ime ụgbọ mmiri zip). Laboratory.The samples e air-Fikiere na-emepụta pulverized samples, pulverized site a n'ibu usoro (Fritsch diski igwe igwe), na sieved (sieve size 2 mm) .Ebe 1 gram nke a mịrị amị, homogenized na sieved ala samples n'ụzọ doro anya kpọrọ teflon bottles. Na onye ọ bụla Teflon arịa, na-ekesa 7 ml nke 65% Hc Hc 35% nke 35% H. onye na-ekesa - otu maka acid ọ bụla), na-ekpuchi obere ọkụ ma kwe ka ihe atụ ahụ guzoro n'otu abalị maka mmeghachi omume (aqua regia program) . Tinye ihe dị elu na efere ígwè na-ekpo ọkụ (okpomọkụ: 100 W na 160 Celsius C) maka 2 h iji mee ka usoro mgbaze nke ihe ndị ahụ dị nro, wee dị jụụ. Nke ahụ, iyo ọnụọgụ ahụ diluted na mmiri PVC PVC na mmiri na-eme ka ọ bụrụ nke ptes (cd, na-ekpebi ya ICP-Oes (Inductively Coupled Plasma Optical Emission Spectroscopy) (Thermo Fisher Scientific, USA) dị ka usoro ọkọlọtọ na nkwekọrịta. Gbaa mbọ hụ na njikwa mma na njikwa (QA / QC) (SRM NIST 2711a Montana II Ala) .PTEs nwere oke nchọpụta n'okpuru ọkara ka ewepụrụ n'ọmụmụ ihe a. Achọpụtara PTE na njedebe nke ọmụmụ a. 0.0004.(gị) .Na mgbakwunye, a na-eme ka njikwa mma na usoro nkwenye dị mma maka nyocha nke ọ bụla site na nyochaa ụkpụrụ ntụaka. Iji hụ na e belatara njehie, a na-eme nyocha abụọ.
Empirical Bayesian Kriging (EBK) bụ otu n'ime ọtụtụ geostatistical interpolation usoro eji na modeling na iche iche ubi dị ka ala sayensị. N'adịghị ka ndị ọzọ kriging interpolation usoro, EBK dị iche na omenala kriging ụzọ site n'ịtụle njehie e mere atụmatụ site na semivariogram model.Na EBK interpolation, ọtụtụ semivariogram ụdị na-agbakọta a semivariogram oge a semivariogram usoro na-agbakọta a semivariogram usoro. mee ka ụzọ maka ejighị n'aka na mmemme metụtara nke a ibé nke semivariogram nke mejupụtara a ukwuu mgbagwoju akụkụ nke a zuru ezu kriging usoro.The interpolation usoro nke EBK na-eso atọ njirisi tụrụ aro site Krivoruchko50, (a) ihe nlereanya na-eme atụmatụ na semivariogram si dataset dataset (b) ọhụrụ buru amụma uru maka onye ọ bụla ntinye dataset ọnọdụ dabeere na n'ịwa nke ikpeazụ dataset ihe nlereanya na-adabere na nke ikpeazụ. simulated dataset.Enyere iwu nha nhata nke Bayesian dị ka azụ
Ebe \ (Prob \ ekpe (A \ nri) \) na-anọchi anya tupu, \ (Prob \ ekpe (B \ nri) \) oke ihe gbasara nke puru omume na-eleghara anya n'ọtụtụ ọnọdụ, \ (Prob (B, A) \ ) .The semivariogram ngụkọta oge dabeere Bayes 'iwu, nke na-egosi na propensity nke chọpụtara datasets na ike ike kere si na semivariogram nke na-ekwu, nke na-ekwu na semivariogram. kedu ka ọ ga-esi mepụta dataset nke nleba anya site na semivariogram.
A support vector igwe bụ igwe mmụta algọridim nke na-emepụta ihe kacha nkewa hyperplane ịmata ọdịiche dị ma ọ bụghị linearly onwe classes.Vapnik51 kere ebumnuche classification algọridim, ma ọ na-adịbeghị anya e ji dozie regression-gbakwasara nsogbu. Dị ka Li et al.52, SVM bụ otu n'ime ndị kasị mma classifier usoro na e jiriwo Machine SVM dị iche iche mpaghara mpaghara. A na-eji Regression - SVMR) mee ihe na nyocha a.Cherkassky na Mulier53 bụ ndị ọsụ ụzọ SVMR dị ka kernel-based regression, nke a na-eme ya site na iji usoro nkwụghachi azụ na-arụ ọrụ na-arụ ọrụ nke ọtụtụ mba. et al. 55, epsilon (ε) -SVMR na-eji dataset a zụrụ azụ nweta ihe nnochite anya dị ka ọrụ epsilon-enweghị mmetụta na-etinye aka na map data n'adabereghị na nke kacha mma epsilon bias site na ọzụzụ na data njikọ. A na-eleghara njehie anya nke preset anya site na uru ahụ n'ezie, ma ọ bụrụ na njehie ahụ dị ukwuu karịa ε (ε), ihe onwunwe ala ahụ na-ebelata ihe mgbagwoju anya nke ihe nleba anya nke ihe nleba anya nke ihe nleba anya nke ihe nleba anya nke ihe omuma nke ihe omuma nke na-eme ka ihe omuma nke ihe omuma ya na-ebelata ihe omuma nke ihe omuma nke ndi ozo na-eme ka ihe omuma nke oma na-eme ka o doo anya na ihe omuma nke ndi ozo di iche iche. vectors. Equation nke Vapnik51 tụrụ aro ka egosiri n'okpuru.
ebe b na-anọchi anya ọnụ ụzọ scalar, \(K\ ekpe ({x}_{,}{ x}_{k}\right)\) na-anọchi anya ọrụ kernel, \(\alpha \) na-anọchi anya Lagrange multiplier, N na-anọchi anya ọnụọgụ dataset, \({x}_{k}\) na-anọchi anya ntinye data, na \(y\) ọrụ bụ data S. bụ Gaussian radial base function (RBF) .A na-etinye kernel RBF iji chọpụta ụdị SVMR kacha mma, nke dị oke mkpa iji nweta ihe kacha ntanye ntaramahụhụ maka C na kernel parameter gamma (γ) maka data ọzụzụ PTE. Nke mbụ, anyị nyochara ọzụzụ ọzụzụ wee nwalee arụmọrụ nlereanya na usoro ntinye aka nke SVMR.
A multiple linear regression model (MLR) bụ regression nlereanya nke na-anọchi anya mmekọrịta dị n'etiti nzaghachi mgbanwe na a ọnụ ọgụgụ nke amụma variables site na iji linear pooled parameters gbakọọ site na iji kasị nta square usoro. mmekọrita na mgbanwe nkọwa.Nha nhata MLR bụ
ebe y bụ mgbanwe nzaghachi, \ (a \) bụ intercept, n bụ ọnụọgụ ndị amụma, \ ({b}_{1}\) bụ akụkụ azụ azụ nke ọnụọgụgụ, \ ({x}_{ i} \) na-anọchite anya amụma ma ọ bụ mgbanwe nkọwa, na \ ({\varepsilon}_{i}\) na-anọchi anya njehie dị na ihe nlereanya ahụ.
A na-enweta ụdị ndị a gwakọtara site na sandwiching EBK na SVMR na MLR. Nke a na-eme site n'iwepụta ụkpụrụ ndị e buru n'amụma site na EBK interpolation. A na-enweta ụkpụrụ ndị a na-enweta site na ca, K, na Mg na-ejikọta ya site na usoro nchịkọta iji nweta mgbanwe ọhụrụ, dị ka CaK, CaMg, na KMg. Ihe ndị na-emepụta ihe bụ CaK, CaMg, na KMg. Ihe nke anọ, CaKM na-agbanwe agbanwe. mgbanwe ndị a nwetara bụ Ca, K, Mg, CaK, CaMg, KMg na CaKMg. Ndị a dị iche iche ghọrọ ndị amụma anyị, na-enyere aka ịkọ ọkwa nickel na obodo ukwu na obodo obodo. A na-eme SVMR algọridim na ndị na-ebu amụma iji nweta ihe atụ agwakọta Empirical Bayesian Kriging-Support Vector Machine (EKK_SVMar) na-enwetakwa ihe nlereanya ML na-agbanwe agbanwe. Empirical Bayesian Kriging-Multiple Linear Regression (EBK_MLR) .Ọtụtụ, mgbanwe Ca, K, Mg, CaK, CaMg, KMg, na CaKMg na-eji dị ka covariates dị ka amụma nke Ni ọdịnaya n'ime obodo mepere emepe na mpụta obodo Ona. The kasị anabata ụdị nke enwetara (EBK_SVM ma ọ bụ EBK_ML ga-arụ ọrụ onwe ya). egosiri ihe omumu a na eserese 2.
Iji SeOM aghọwo ngwá ọrụ na-ewu ewu maka ịhazi, nyochaa, na ịkọ amụma data na mpaghara ego, ahụike, ụlọ ọrụ, ọnụ ọgụgụ, sayensị ala, na ndị ọzọ.SeOM na-emepụta site na iji netwọk neural na-enweghị nlekọta na usoro mmụta na-enweghị nlekọta maka nhazi, nyocha, na amụma. N'ime ọmụmụ ihe a, a na-eji SeOM na-ahụ anya na nhazi nke Nini dabere na nhazi nke obodo na-adabere na usoro nhazi nke obodo na-atụ anya na nhazi nke Nii. A na-eji nyocha SeOM mee ihe dị ka n ntinye-akụkụ vector variables43,56.Melssen et al. 57 na-akọwa njikọ nke vector ntinye n'ime netwọk neural site na otu ntinye oyi akwa gaa na vector mmepụta nwere otu vector dị arọ. Ihe mmepụta nke SeOM na-emepụta bụ maapụ akụkụ abụọ nke nwere neurons dị iche iche ma ọ bụ ọnụ na-etinye n'ime maps hexagonal, okirikiri ma ọ bụ square topological dị ka nso ha si dị. Comparing map sizes based on error topographic (Q TE) nlereanya na njehie topographic metric (QTE) quantization. 0.086 na 0.904, n'otu n'otu, a na-ahọrọ, nke bụ 55-map unit (5 × 11) . A na-ekpebi nhazi nke neuron dị ka ọnụ ọgụgụ nke ọnụ ọgụgụ dị na nkwekọ nhụsianya.
Ọnụ ọgụgụ nke data e ji mee ihe n'ọmụmụ ihe a bụ 115 samples. A na-eji usoro a na-enweghị usoro iji kewaa data n'ime data ule (25% maka nkwenye) na nhazi data ọzụzụ (75% maka nhazi) .A na-eji dataset ọzụzụ mee ihe n'ịmepụta ihe ngosi regression (calibration), na nyocha dataset na-eji iji nyochaa ikike izugbe58. Nke a mere iji nyochaa ịdị mma nke ụdị dị iche iche nke nickel gara amụma. Usoro nkwenye okpukpu iri nke okpukpu iri, ugboro ise ugboro ise. A na-eji mgbanwe ndị a na-emepụta site na EBK interpolation dị ka ndị amụma ma ọ bụ nkọwa nkọwa iji kọwaa ihe mgbaru ọsọ (PTE) . A na-edozi ihe nlere anya na RStudio site na iji nchịkọta ngwugwu (Kohonen), ụlọ akwụkwọ (nlekọta), ụlọ akwụkwọ (modelr), ụlọ akwụkwọ ("e1071"), ụlọ akwụkwọ ("e1071"), ụlọ akwụkwọ ("plyr" ), ụlọ akwụkwọ ("plyr" ), ụlọ akwụkwọ ("plyr"), ụlọ akwụkwọ ("plyr"), ụlọ akwụkwọ ("plyr"), ụlọ akwụkwọ ("plyr") ("Metrics").
A na-eji paramita nkwado dị iche iche iji chọpụta ihe nlereanya kachasị mma maka ịkọ ọkwa nickel na ala na iji nyochaa izi ezi nke ihe nlereanya ahụ na nkwenye ya. A na-enyocha ụdị hybridization site na iji njehie zuru oke (MAE), mgbọrọgwụ pụtara square njehie (RMSE), na R-squared ma ọ bụ ọnụọgụ ọnụọgụ (R2) .RMSE na-akọwa ọdịiche nke oke na azịza na regression model. jikoro na-akọwa ike amụma nke ihe nlereanya ahụ, ebe MAE na-ekpebi ọnụ ahịa ọnụ ọgụgụ n'ezie. Uru R2 ga-adị elu iji nyochaa ihe nlereanya ngwakọta kachasị mma site na iji akara nkwado, na-eru nso na 1, nke dị elu nke ziri ezi. Dị ka Li et al. 59, uru R2 nke 0.75 ma ọ bụ karịa ka a na-ewere dị ka ezigbo amụma; site na 0.5 ruo 0.75 bụ ihe nlereanya a na-anabata nke ọma, na n'okpuru 0.5 bụ arụmọrụ nlereanya na-adịghị anabata. Mgbe ị na-ahọrọ ihe nlereanya site na iji usoro nyocha nke RMSE na MAE, ụkpụrụ ndị dị ala enwetara zuru oke ma weere ya dị ka nhọrọ kacha mma. Ngụkọta na-esonụ na-akọwa usoro nkwenye.
ebe n na-anọchi anya nha nke uru hụrụ\({Y}_{i}\) na-anọchi anya nzaghachi atụnyere, na \({\ widehat{Y}}_{i}\) na-anọchikwa anya uru nzaghachi e buru n'amụma, ya mere, maka nke mbụ i nhụbanya.
A na-egosiputa nkọwa ndekọ ọnụ ọgụgụ nke amụma amụma na mgbanwe nzaghachi na Tebụl 1, na-egosi pụtara, ọkọlọtọ deviation (SD), ọnụọgụ nke mgbanwe (CV), kacha nta, kacha, kurtosis, na skewness. The kacha nta na kacha ụkpụrụ nke ihe ndị dị na mbelata usoro nke Mg N'ihi ihe dị iche iche a tụrụ atụ nke ihe ndị a tụrụ atụ, nkesa nkesa data nke ihe ndị ahụ na-egosipụta skewness dị iche iche. Ntugharị na kurtosis nke ihe dị iche iche sitere na 1.53 ruo 7.24 na 2.49 ruo 54.16, n'otu n'otu. Ihe niile gbakọrọ nwere skewness na kurtosis na-ekesa n'elu +1, na-eme ka data dị n'elu nkesa na-eme ka ọ dị elu +1. ziri ezi na peaked.Cv ndị a na-eme atụmatụ nke ihe ndị ahụ na-egosikwa na K, Mg, na Ni na-egosipụta mgbanwe dị oke oke, ebe Ca nwere oke mgbanwe dị elu. CV nke K, Ni na mg na-akọwa nkesa otu ha. Ọzọkwa, nkesa Ca bụ nke na-abụghị uwe na isi mmalite nwere ike imetụta ọkwa nkwalite ya.
Njikọ nke mgbanwe ndị na-ebu amụma na ihe ndị na-azaghachi na-egosi njikọ dị mma n'etiti ihe ndị ahụ (lee foto 3) . Mmekọrịta ahụ gosiri na CaK gosipụtara mmekọrịta dị oke ọnụ na uru r = 0.53, dị ka CaNi. Ọ bụ ezie na Ca na K na-egosi mkpakọrịta dị umeala n'obi na ibe ha, ndị nchọpụta dị ka Kingston et al. 68 na Santo69 na-atụ aro na ọkwa ha dị na ala dị n'ụzọ ziri ezi. Otú ọ dị, Ca na Mg na-emegide K, ma CaK na-ejikọta ya nke ọma. Nke a nwere ike ịbụ n'ihi ntinye nke fatịlaịza dị ka potassium carbonate, nke dị 56% dị elu na potassium. Potassium na-ejikọta ya na magnesium (KM r = 0.63), n'ihi na potassium fatịlaịza na-ejikọta ya na abụọ magnesium fatịlaịza. nitrate, na potash na-etinye aka na ala iji mee ka ọkwa ha dịkwuo elu. Nickel na-ejikọta ya na Ca, K na Mg nke ọma na ụkpụrụ r = 0.52, 0.63 na 0.55. Mmekọrịta ndị metụtara calcium, magnesium, na PTE dị ka nickel dị mgbagwoju anya, ma ka o sina dị, ma magnesium na calcium na-ebelata mmetụta nke calcium na calcium absorption. mmetụta nsi nke nickel na ala.
Matrix njikọ maka ihe ndị na-egosi mmekọrịta dị n'etiti ndị amụma na nzaghachi (Rịba ama: ọnụ ọgụgụ a na-agụnye nkwụsịtụ n'etiti ihe ndị dị mkpa, ọkwa dị mkpa dabeere na p <0,001).
Ọnụ ọgụgụ 4 na-egosipụta nkesa nkesa nke ihe dị iche iche. Dị ka Burgos et al70 si kwuo, ntinye nke nkesa mbara igwe bụ usoro eji akọwapụta ma gosipụta ebe dị ọkụ na mpaghara ndị rụrụ arụ. A na-ahụ ọkwa nkwalite nke Ca na Fig. quicklime (calcium oxide) iji belata ala acidity na iji ígwè igwe igwe dị ka alkaline oxygen na steelmaking usoro. N'aka nke ọzọ, ndị ọrụ ugbo ndị ọzọ na-ahọrọ iji calcium hydroxide na acidic Ona na neutralize pH, nke na-abawanye calcium ọdịnaya nke ala71.Potassium na-egosikwa na-ekpo ọkụ tụrụ na n'ebe ugwu ọdịda anyanwụ na n'ebe ọwụwa anyanwụ nke map.The Northwest obodo nwere ike na-mere ka ụkpụrụ nke calcium hydroxide na acidic ala neutralize pH. NPK na ngwa ngwa potash.Nke a kwekọrọ na ọmụmụ ihe ndị ọzọ, dị ka Madaras na Lipavský72, Madaras et al.73, Pulkrabova et al.74, Asare et al.75, bụ ndị chọpụtara na nkwụsi ike ala na ọgwụgwọ na KCl na NPK mere ka ọdịnaya K dị elu na ala. Ọganihu Potassium gbasara ohere na ugwu ọdịda anyanwụ nke map nkesa nwere ike ịbụ n'ihi iji fatịlaịza dabeere na potassium dị ka potassium chloride, potassium sulfate, potassium nitrate, potash, na potash iji mee ka ọdịnaya potassium dị n'ime ala dara ogbenye.Zádorová et al. 76 na Tlustoš et al. 77 kwuputara na ntinye nke fatịlaịza sitere na K na-amụba ọdịnaya K n'ime ala ma na-eme ka ihe oriri na-edozi ahụ dịkwuo elu n'ime ogologo oge, karịsịa K na Mg na-egosi ebe dị ọkụ n'ime ala. Ọkụ na-ekpo ọkụ na-agafeghị oke na ugwu ọdịda anyanwụ nke map na n'ebe ndịda ọwụwa anyanwụ nke map. Colloidal fixation na ala na-eme ka ntinye nke magnesium n'ime ala na-eme ka ọ ghara ịdị na-eme ka ịta ahụhụ nke magnesium na ala na-acha odo odo. chlorosis.Magnesium dabeere na fatịlaịza, dị ka potassium magnesium sulfate, magnesium sulfate, na Kieserite, na-emeso erughị eru (osisi na-egosi na-acha odo odo, red, ma ọ bụ agba aja aja, na-egosi magnesium erughi) na ala na a nkịtị pH range6.The mkpokọta nke nickel na obodo mepere emepe na mpụta-obodo ukwu ala na-ebupụta nwere ike ịbụ n'ihi na anthropogenic mmepụta nke anthropogenic ọrụ dị ka agriculture na-adịghị mkpa.
Nkesa ihe gbasara mbara ala [e ji ArcGIS Desktop mepụta maapụ nkesa oghere (ESRI, Inc, Version 10.7, URL: https://desktop.arcgis.com).]
N'aka nke ọzọ, RMSE na MAE nke Ni dị nso na zero (0.86 RMSE, -0.08 MAE) .N'aka nke ọzọ, ma RMSE na MAE ụkpụrụ nke K na-anabata .RMSE na MAE. Nsonaazụ dị ukwuu maka calcium na magnesium na ihe dị iche iche nke MAE. MAE nke ọmụmụ a na-eji EBK na-ebu amụma Ni ka achọpụtara dị mma karịa nsonaazụ John et al. 54 na-eji synergistic kriging na-ebu amụma gbasara mkpokọta S na ala site na iji otu data anakọtara. Ihe EBK anyị mụrụ na-ejikọta na nke Fabijaczyk et al. 41, Yan et al. 79, Beguin et al. 80, Adhikary et al. 81 na John et al. 82, ọkachasị K na Ni.
A na-enyocha arụmọrụ nke ụzọ onye ọ bụla maka ịkọ ọdịnaya nickel n'ime obodo na obodo dịpụrụ adịpụ site na iji arụmọrụ nke ụdị (Table 3) .Nkwenye ihe nlereanya na nyocha ziri ezi kwadoro na Ca_Mg_K amụma jikọtara ya na EBK SVMR nlereanya na-arụpụta ihe kasị mma.Calibration nlereanya Ca_Mg_K-EBK_SVMR nlereanya ab. 0.637 (R2), 95.479 mg / kg (RMSE) na 77.368 mg / kg (MAE) Ca_Mg_K-SVMR bụ 0.663 (R2), 235.974 mg / kg (RMSE) na 166.946 mg / kg (erughị MAE) bụ ezigbo RN2 uru. Ca_Mg_K-SVMR (0.663 mg/kg R2) na Ca_Mg-EBK_SVMR (0.643 = R2); nsonaazụ RMSE na MAE ha dị elu karịa nke Ca_Mg_K-EBK_SVMR (R2 0.637) (lee Isiokwu 3) . Tụkwasị na nke ahụ, RMSE na MAE nke Ca_Mg-EBK_SVMR (RMSE = 1664.64 na MAE = 1031.49) na-asọpụrụ 137.5, na ndị na-asọpụrụ 137.5. Ca_Mg_K-EBK_SVMR.N'otu aka ahụ, RMSE na MAE nke Ca_Mg-K SVMR (RMSE = 235.974 na MAE = 166.946) nlereanya bụ 2.5 na 2.2 buru ibu karịa nke Ca_Mg_K-EBK_SVMR na-egosi otú RMSE si gbakọọ na nsonaazụ ya. na ahịrị kacha mma. A hụrụ RSME dị elu na MAE. Dị ka Kebonye et al. 46 na John et al. 54, ka RMSE na MAE dị nso na efu, ihe ka mma.SVMR na EBK_SVMR nwere ọnụ ọgụgụ dị elu nke RSME na MAE. Achọpụtara na atụmatụ RSME na-anọgide na-adị elu karịa ụkpụrụ MAE, na-egosi ọnụnọ nke outliers. Dị ka Legates na McCabe83 si kwuo, njedebe SE na-atụ aro ka ọ gafere oke (njedebe) dị ka ihe na-eme ka ọ bụrụ ihe na-agafe agafe. ihe na-egosi ọnụnọ nke outliers. Nke a pụtara na ndị ọzọ heterogeneous na dataset, na elu MAE na RMSE ụkpụrụ. The izi ezi nke cross-validation ntule nke Ca_Mg_K-EBK_SVMR agwakọta nlereanya maka ịkọ Ni ọdịnaya n'ime obodo mepere emepe na ala dịpụrụ adịpụ bụ 63.70% dị ka Li et al. 59, ọkwa a nke ziri ezi bụ ihe nlereanya a na-anabata nke ọma. A na-atụnyere nsonaazụ dị ugbu a na ọmụmụ ihe gara aga nke Tarasov et al. 36 onye ngwakọ ụdị kere MLPRK (Multilayer Perceptron Residual Kriging), metụtara EBK_SVMR ziri ezi nwale index kọrọ na nke ugbu a ọmụmụ, RMSE (210) na The MAE (167.5) bụ elu karịa anyị pụta na nke ugbu a ọmụmụ (RMSE 95.479, MAE 77.368 nke ugbu a ọmụmụ). (0.637) na nke Tarasov et al. 36 (0.544), o doro anya na ọnụọgụ nke mkpebi siri ike (R2) dị elu na ụdị a gwakọtara ọnụ. The oke nke njehie (RMSE na MAE) (EBK SVMR) maka agwakọta nlereanya bụ ugboro abụọ ala. N'otu aka ahụ, Sergeev et al.34 dekọrọ 0.28 (R2) maka mepụtara ngwakọ nlereanya e dekọrọ (Multilayer KR2). 0.637 (R2) .Ọkwa amụma ziri ezi nke ihe nlereanya a (EBK SVMR) bụ 63.7%, ebe amụma amụma Sergeev et al. 34 bụ 28%.Map ikpeazụ (Fig 5) nke e kere site na iji ụdị EBK_SVMR na Ca_Mg_K dị ka onye amụma na-egosi amụma nke ebe dị ọkụ ma na-agafe agafe na nickel n'elu ebe ọmụmụ ihe dum. Nke a pụtara na ntinye nke nickel na ebe ọmụmụ ihe na-abụkarị agafeghị oke, na-enwe oke dị elu na mpaghara ụfọdụ.
A na-anọchi anya maapụ amụma ikpeazụ site na iji ụdị ngwakọ EBK_SVMR yana iji Ca_Mg_K dị ka onye amụma.
Egosiputara na onu ogugu 6 bu ihe omuma nke PTE dika ihe eji eme ihe nke nwere neurons n'otu n'otu. Ọ dịghị nke ụgbọ elu ndị na-emepụta ihe na-egosipụta otu ụkpụrụ agba dị ka egosipụtara. Otú ọ dị, ọnụ ọgụgụ kwesịrị ekwesị nke neurons kwa eserese eserese bụ 55. SeOM na-emepụta site na iji ụdị dị iche iche nke agba, na ihe yiri nke agba agba, na-atụnyere ihe onwunwe nke samples. Acorder ha nwere ike ime ka ihe ndị yiri ya, na-eme ka ihe ndị yiri ya na ihe atụ nke K. agba ụkpụrụ ka otu elu neurons na ọtụtụ ala neurons.Ya mere, CaK na CaMg na-ekerịta ụfọdụ myirịta na nnọọ elu-usoro neurons na ala-na-agafeghị oke na agba ụkpụrụ.Ma ụdị amụma amụma ịta nke Ni na ala site n'igosipụta ọkara na elu hues nke na agba dị ka red, oroma na yellow.The KMg nlereanya na-egosiputa ọtụtụ elu agba ụkpụrụ dabere na obere agba na-adabere na nkenke agba na-adabere na ọnụ ọgụgụ dị ala proportci. elu, usoro nkesa nkesa nke akụkụ nke ihe nlereanya ahụ gosipụtara ụkpụrụ agba agba dị elu nke na-egosi ike ịta ahụhụ nke nickel na ala (lee Figure 4) .Ụgbọ elu nke ihe nlereanya CakMg na-egosi ụdị agba dị iche iche site na ala ruo elu dị ka ihe ziri ezi agba. Ntụle nke nickel n'ime ime obodo na ala ime obodo. Foto nke 7 na-egosi usoro contour na k-pụtara na-achịkọta na maapụ, kewara n'ime ụyọkọ atọ dabere na uru e buru n'amụma na ụdị nke ọ bụla. Usoro nhazi na-anọchi anya ọnụ ọgụgụ kachasị mma nke ụyọkọ. N'ime ihe nlele ala 115 anakọtara, ụdị 1 nwetara ihe nlele nke ala, ụyọkọ 3 3 C nwetara ihe nlele nke ala, ụyọkọ 3. natara ihe nlele 8. A na-eme ka ngwakọta nke atụmatụ atụmatụ atụmatụ asaa dị mfe iji nye ohere maka nkọwa ụyọkọ ziri ezi. N'ihi ọtụtụ usoro anthropogenic na eke na-emetụta nhazi ala, ọ na-esiri ike inwe ụdị ụyọkọ dị iche iche nke ọma na map78 nke SeOM kesara.
Mbupụta ụgbọ elu nke ọ bụla Empirical Bayesian Kriging Support Vector Machine (EBK_SVM_SeOM) na-agbanwe.[Ejiri maapụ SeOM site na iji RStudio (ụdị 1.4.1717: https://www.rstudio.com/).]
Akụkụ nkewa ụyọkọ dị iche iche [Ejiri maapụ SeOM site na iji RStudio (ụdị 1.4.1717: https://www.rstudio.com/)]
Ọmụmụ ihe a na-egosi n'ụzọ doro anya usoro nhazi maka nhazi nickel n'ime obodo na obodo dịpụrụ adịpụ. Ihe ọmụmụ ahụ nwalere usoro nhazi dị iche iche, na-ejikọta ihe na usoro nhazi, iji nweta ụzọ kachasị mma iji buru amụma gbasara nickel n'ime ala. akwado planar spatial nkesa nke components exhibited by EBK_SVMR (lee Figure 5) .The results na-egosi na support vector igwe regression nlereanya (Ca Mg K-SVMR) amụma ịta nke Ni na ala dị ka otu nlereanya, ma nkwado na ziri ezi nwale parameters na-egosi nnọọ elu njehie n'usoro nke RMSE na MAE.On n'aka nke ọzọ, na-arụ ọrụ na-modeling ML n'ọrụ na-eṅomi n'ihi na EBR na-eme ihe nlereanya na-eme ihe nlereanya na-eme ihe nlereanya. ọnụ ala dị ala nke ọnụọgụ nke mkpebi siri ike (R2) .Enwetara nsonaazụ dị mma site na iji EBK SVMR na ihe ndị jikọtara ọnụ (CaKMg) na obere RMSE na njehie MAE dị ala na nke ziri ezi nke 63.7% .Ọ na-atụgharị na ijikọta EBK algọridim na igwe mmụta algọridim nwere ike ịmepụta algorithm ngwakọ nke nwere ike ịkọ amụma ntinye nke PTEs na ala. Nsonaazụ Ni K nwere ike imeziwanye ihe na-eme ka nyocha nke Ca n'ime ala. nke Ni na ala.Nke a pụtara na na-aga n'ihu ngwa nke nickel dabeere fatịlaịza na ulo oru mmetọ nke ala site nchara ụlọ ọrụ nwere ọchịchọ dịkwuo ịta nke nickel n'ime ala.Nke a na-amụ gosiri na EBK nlereanya nwere ike ibelata larịị nke njehie na melite ziri ezi nke nlereanya nke ala gbasara ohere nkesa n'ime obodo ma ọ bụ mpụta obodo Ona. N'ozuzu, anyị na-atụ aro na-etinye n'ime ala PEB TEK nlereanya. na mgbakwunye, anyị na-atụ aro iji EBK ka hybridize dị iche iche igwe mmụta algọridim.Ni ịta e buru amụma iji ọcha dị ka covariates; Otú ọ dị, iji ndị ọzọ covariates ga-eme ka arụmọrụ nke ihe nlereanya ahụ dịkwuo mma, nke a pụrụ iwere dị ka njedebe nke ọrụ ugbu a. Mbelata ọzọ nke ọmụmụ ihe a bụ na ọnụ ọgụgụ nke datasets bụ 115. Ya mere, ọ bụrụ na a na-enyekwu data, a pụrụ imeziwanye arụmọrụ nke usoro hybridization nke a na-atụ aro.
PlantProbs.net.Nickel na Osisi na ala https://plantprobs.net/plant/nutrientsImbalances/sodium.html (Nwetara 28 Eprel 2021).
Kasprzak, KS Nickel na-aga n'ihu na toxicology gburugburu ebe obibi ọgbara ọhụrụ.surroundings.toxicology.11, 145-183 (1987).
Cempel, M. & Nikel, G. Nickel: Nyochaa isi mmalite ya na toxicology gburugburu ebe obibi. Polish J. Environment.Stud.15, 375-382 (2006).
Freedman, B. & Hutchinson, TC ndenye mmetọ sitere na ikuku na mkpokọta n'ime ala na ahịhịa n'akụkụ ebe mgbaze nickel-copper na Sudbury, Ontario, Canada.can.J. Bot.58 (1), 108-132.https://doi.org/10.1139/b80-014 (1980).
Manyiwa, T. et al. Ọla dị arọ na ala, osisi na ihe ize ndụ jikọtara ya na anụ ọhịa na-ata nri na nso Selebi-Phikwe ọla kọpa-nickel mine na Botswana.surroundings.Geochemistry.Health https://doi.org/10.1007/s10653-021-02021x.
Cabata-Pendias.Kabata-Pendias A. 2011. Chọpụta ihe dị na ala na… - Google Ọkà mmụta https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=Kabata-Pendias+A.+2011.+Trace+ Elements+n'ime+ala+na+osisi.+4th+ed.+New+York+%28NY%29%3A+CRC+Press&btnG= (Nwetara 24 Nov 2020).
Almås, A., Singh, B., Agriculture, TS-NJ nke & 1995, undefined.Mmetụta nke ụlọ ọrụ nickel Russia na oke ígwè dị arọ na ala ugbo na ahịhịa na Soer-Varanger, Norway.agris.fao.org.
Nielsen, GD et al.Nickel absorption na njigide na mmiri ọṅụṅụ metụtara nri nri na nickel sensitivity.toxicology.application.Pharmacodynamics.154, 67-75 (1999).
Costa, M. & Klein, CB Nickel carcinogenesis, mmụpụta, epigenetics ma ọ bụ nhọrọ. gburugburu. Health Perspective.107, 2 (1999).
Ajman, PC; Ajado, SK; Borůvka, L.; Bini, JKM; Sarkody, VYO; Cobonye, ​​NM; Usoro nyocha nke ihe ndị nwere ike igbu egbu: nyocha bibliometric.Environmental Geochemistry and Health.Springer Science & Business Media BV 2020.https://doi.org/10.1007/s10653-020-00742-9.
Minasny, B. & McBratney, AB Digital Ala Map: Akụkọ dị nkenke na nkuzi ụfọdụ.Geoderma 264, 301–311.https://doi.org/10.1016/j.geoderma.2015.07.017 (2016).
McBratney, AB, Mendonça Santos, ML & Minasny, B. Na nkewa ala dijitalụ.Geoderma 117 (1-2), 3-52.https://doi.org/10.1016/S0016-7061(03)00223-4 (2003).
Deutsch.CV Geostatistical Reservoir Modeling,… – Google Ọkà mmụta https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=CV+Deutsch%2C+2002%2C+Geostatistical+Reservoir+Modeling%2C +Oxford+University+Press%2C+376+ibe.+&btnG= (Nwetara 28 Eprel 2021).