{"id":41221,"date":"2020-02-06T20:42:46","date_gmt":"2020-02-06T17:42:46","guid":{"rendered":"https:\/\/prohoster.info\/blog\/blog_prohoster\/python-gateway-v-intersystems-iris"},"modified":"2020-02-06T20:42:46","modified_gmt":"2020-02-06T17:42:46","slug":"python-gateway-v-intersystems-iris","status":"publish","type":"post","link":"https:\/\/prohoster.info\/sq\/blog\/python-gateway-v-intersystems-iris","title":{"rendered":"Python Gateway n\u00eb InterSystems IRIS","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

Ky kjo artikull i kushtohet Python Gateway \u2014 nj\u00eb projekt komuniteti me kod t\u00eb hapur p\u00ebr platform\u00ebn e t\u00eb dh\u00ebnave InterSystems IRIS. Ky projekt lejon orkestrimin e \u00e7do algoritmi t\u00eb m\u00ebsimit t\u00eb makineris\u00eb, i krijuar n\u00eb gjuh\u00ebn Python (mjedisi kryesor p\u00ebr shum\u00eb Shkenc\u00ebtar\u00eb t\u00eb T\u00eb Dh\u00ebnave), duke p\u00ebrdorur bibliotekat e shumta t\u00eb gatshme p\u00ebr t\u00eb ndihmuar n\u00eb krijimin e zgjidhjeve adaptuese, robotizuese analitike AI\/ML n\u00eb platform\u00ebn InterSystems IRIS. N\u00eb k\u00ebt\u00eb artikull, do t\u00eb tregoj se si InterSystems IRIS mund t\u00eb orkestroj\u00eb proceset n\u00eb gjuh\u00ebn Python, t\u00eb realizoj\u00eb me efektivitet transferimin dyansh\u00ebm t\u00eb t\u00eb dh\u00ebnave dhe t\u00eb krijoj\u00eb procese inteligjente biznesi.<\/p>\n

<\/a><\/noindex><\/p>\n

Plani<\/h1>\n

<\/p>\n

    \n
  1. Hyrje.<\/li>\n
  2. Mjetet.<\/li>\n
  3. Instalimi.<\/li>\n
  4. API.<\/li>\n
  5. Nd\u00ebroperabiliteti.<\/li>\n
  6. Jupyter Notebook.<\/li>\n
  7. P\u00ebrfundime.<\/li>\n
  8. Lidhjet.<\/li>\n
  9. MLToolkit.<\/li>\n<\/ol>\n

    <\/p>\n

    Hyrje<\/h1>\n

    <\/p>\n

    Python \u00ebsht\u00eb nj\u00eb gjuh\u00eb programimi n\u00eb nivel t\u00eb lart\u00eb, p\u00ebr q\u00ebllime t\u00eb p\u00ebrgjithshme, e orientuar drejt rritjes s\u00eb produktivitetit t\u00eb zhvilluesit dhe lexueshm\u00ebris\u00eb s\u00eb kodit. N\u00eb k\u00ebt\u00eb seri artikujsh, do t\u00eb flas p\u00ebr mund\u00ebsit\u00eb e p\u00ebrdorimit t\u00eb gjuh\u00ebs Python n\u00eb platform\u00ebn InterSystems IRIS, duke qen\u00eb se fokusi kryesor i k\u00ebtij artikulli \u00ebsht\u00eb aplikimi i Python si gjuh\u00eb p\u00ebr krijimin dhe p\u00ebrdorimin e modeleve t\u00eb m\u00ebsimit t\u00eb makineris\u00eb.<\/p>\n

    <\/p>\n

    M\u00ebsimi i makineris\u00eb (ML) \u00ebsht\u00eb nj\u00eb klas\u00eb metodash t\u00eb inteligjenc\u00ebs artificiale, me karakteristik\u00ebn e t\u00eb cil\u00ebs \u00ebsht\u00eb jo zgjidhja e drejtp\u00ebrdrejt\u00eb e problemit, por m\u00ebsimi gjat\u00eb zgjidhjes s\u00eb shum\u00eb problemeve t\u00eb ngjashme. <\/p>\n

    <\/p>\n

    Algoritmet dhe modelet e m\u00ebsimit t\u00eb makineris\u00eb po b\u00ebhen gjithnj\u00eb e m\u00eb t\u00eb zakonshme. Ka shum\u00eb arsye p\u00ebr k\u00ebt\u00eb, por gjith\u00e7ka reduktohet n\u00eb aksesueshm\u00ebrin\u00eb, thjesht\u00ebsin\u00eb dhe arritjen e rezultateve praktike. A \u00ebsht\u00eb klasterizimi ose madje modelimi me rrjete nervore nj\u00eb teknologji e re? <\/p>\n

    <\/p>\n

    Sigurisht q\u00eb jo, por tani nuk ka nevoj\u00eb t\u00eb shkruhen qindra mij\u00ebra rreshta kodi p\u00ebr t\u00eb lan\u00e7uar nj\u00eb model, dhe shpenzimet p\u00ebr krijimin dhe p\u00ebrdorimin e modeleve po b\u00ebhen gjithnj\u00eb e m\u00eb t\u00eb vogla.<\/p>\n

    <\/p>\n

    Mjetet po evoluojn\u00eb \u2014 ndon\u00ebse nuk kemi mjete t\u00eb plota t\u00eb orientuara nga GUI p\u00ebr AI\/ML, progresi q\u00eb kemi par\u00eb me shum\u00eb klasa t\u00eb tjera t\u00eb sistemeve informative, p\u00ebr shembull, BI (nga shkrimi i kodit deri te p\u00ebrdorimi i kornizave dhe zgjidhjeve me konfigurim t\u00eb orientuar nga GUI), po ndiqet gjithashtu n\u00eb mjetet p\u00ebr krijimin e AI\/ML. Ne tashm\u00eb kemi kaluar faz\u00ebn e shkruarjes s\u00eb kodit dhe sot p\u00ebrdorim korniza p\u00ebr nd\u00ebrtimin dhe trajnimin e modeleve.<\/p>\n

    <\/p>\n

    P\u00ebrmir\u00ebsime t\u00eb tjera, si\u00e7 \u00ebsht\u00eb mund\u00ebsia p\u00ebr t\u00eb shp\u00ebrndar\u00eb nj\u00eb model t\u00eb paratrainuar, kur p\u00ebrdoruesi p\u00ebrfundimtar duhet thjesht t\u00eb p\u00ebrfundoj\u00eb trajnimin e modelit me t\u00eb dh\u00ebnat e tij specifike, gjithashtu leht\u00ebsojn\u00eb fillimin e p\u00ebrdorimit t\u00eb m\u00ebsimit t\u00eb makineris\u00eb. K\u00ebto arritje e b\u00ebjn\u00eb shum\u00eb m\u00eb t\u00eb leht\u00eb studimin e m\u00ebsimit t\u00eb makineris\u00eb si p\u00ebr specialist\u00ebt, ashtu edhe p\u00ebr kompanit\u00eb n\u00eb t\u00ebr\u00ebsi.<\/p>\n

    <\/p>\n

    Nga ana tjet\u00ebr, ne po mbledhim gjithnj\u00eb e m\u00eb shum\u00eb t\u00eb dh\u00ebna. Fal\u00eb nj\u00eb platforme t\u00eb unifikuar t\u00eb t\u00eb dh\u00ebnave si InterSystems IRIS, t\u00eb gjitha k\u00ebto informacione mund t\u00eb p\u00ebrgatiten menj\u00ebher\u00eb dhe t\u00eb p\u00ebrdoren si t\u00eb dh\u00ebna fillestare p\u00ebr modelet e m\u00ebsimit t\u00eb makineris\u00eb.<\/p>\n

    <\/p>\n

    Me kalimin n\u00eb cloud, nisja e projekteve AI\/ML b\u00ebhet m\u00eb e leht\u00eb se kurr\u00eb. Ne mund t\u00eb konsumohemi vet\u00ebm p\u00ebr burimet q\u00eb na nevojiten. M\u00eb tep\u00ebr, fal\u00eb paralelizimit q\u00eb ofrojn\u00eb platformat cloud, ne mund t\u00eb kursejm\u00eb koh\u00ebn e shpenzuar.<\/p>\n

    <\/p>\n

    Por \u00e7far\u00eb ndodh me rezultatet? K\u00ebtu gjith\u00e7ka b\u00ebhet m\u00eb e komplikuar. Ekzistojn\u00eb shum\u00eb mjete p\u00ebr nd\u00ebrtimin e modeleve, p\u00ebr t\u00eb cilat do t\u00eb flas m\u00eb posht\u00eb. Nd\u00ebrtimi i nj\u00eb modeli t\u00eb mir\u00eb nuk \u00ebsht\u00eb e leht\u00eb, por \u00e7far\u00eb vjen m\u00eb pas? T\u00eb nxjerrim fitime nga p\u00ebrdorimi i modelit nga biznesi gjithashtu \u00ebsht\u00eb nj\u00eb detyr\u00eb e thjesht\u00eb. Rr\u00ebnj\u00ebt e problemit q\u00ebndrojn\u00eb n\u00eb ndarjen e ngarkesave analitike dhe transaksionale, dhe modeleve t\u00eb t\u00eb dh\u00ebnave. Kur e st\u00ebrvitim nj\u00eb model, ne zakonisht e b\u00ebjm\u00eb k\u00ebt\u00eb me t\u00eb dh\u00ebna historike. Por vendi p\u00ebr modelin e nd\u00ebrtuar \u00ebsht\u00eb n\u00eb p\u00ebrpunimin transaksional t\u00eb t\u00eb dh\u00ebnave. \u00c7far\u00eb dobi ka modeli m\u00eb i mir\u00eb p\u00ebr zbulimin e transaksioneve mashtruese, n\u00ebse e aktivizojm\u00eb at\u00eb nj\u00eb her\u00eb n\u00eb dit\u00eb? Mashtruesit jan\u00eb larguar me parat\u00eb. Na nevojitet t\u00eb trajnojm\u00eb modelin me t\u00eb dh\u00ebna historike, por gjithashtu duhet ta aplikojm\u00eb n\u00eb koh\u00eb reale me t\u00eb dh\u00ebna t\u00eb reja q\u00eb vijn\u00eb, n\u00eb m\u00ebnyr\u00eb q\u00eb proceset tona biznesi t\u00eb mund t\u00eb veprojn\u00eb n\u00eb p\u00ebrputhje me parashikimet e b\u00ebra nga modeli.<\/p>\n

    <\/p>\n

    ML Toolkit \u00ebsht\u00eb nj\u00eb paket\u00eb mjetesh, q\u00eb ka si q\u00ebllim pik\u00ebrisht k\u00ebt\u00eb: t\u00eb bashkoj\u00eb modelet dhe mjedisin transaksional, n\u00eb m\u00ebnyr\u00eb q\u00eb modelet e nd\u00ebrtuara t\u00eb mund t\u00eb p\u00ebrdoren leht\u00ebsisht direkt n\u00eb proceset tuaja biznesore. Python Gateway \u00ebsht\u00eb nj\u00eb pjes\u00eb e ML Toolkit dhe siguron integrimin me gjuh\u00ebn Python (ashtu si R Gateway, q\u00eb \u00ebsht\u00eb pjes\u00eb e ML Toolkit, siguron integrimin me gjuh\u00ebn R).<\/p>\n

    <\/p>\n

    Mjetet<\/h1>\n

    <\/p>\n

    Para se vazhdojm\u00eb, do doja t\u00eb p\u00ebrshkruaja disa mjete dhe biblioteka p\u00ebr Python q\u00eb do t\u00eb p\u00ebrdorim m\u00eb von\u00eb.<\/p>\n

    <\/p>\n

    Teknologjit\u00eb<\/h2>\n

    <\/p>\n