{"id":5099,"date":"2023-12-22T16:09:09","date_gmt":"2023-12-22T13:09:09","guid":{"rendered":"https:\/\/datakapital.com\/blog\/?p=5099"},"modified":"2025-08-14T18:03:08","modified_gmt":"2025-08-14T15:03:08","slug":"bist-tahminlemede","status":"publish","type":"post","link":"https:\/\/datakapital.com\/blog\/bist-tahminlemede\/","title":{"rendered":"Bist Tahminlemede Teknik \u0130ndikat\u00f6rlerin \u00d6znitelik \u0130ncelemesi"},"content":{"rendered":"

Ortalama bir insan\u0131n borsaya olan ilgisi ge\u00e7ti\u011fimiz y\u0131llarda \u00fcssel olarak artm\u0131\u015ft\u0131r [1]. Yani herg\u00fcn milyarlarca dolar de\u011ferinde varl\u0131k kar elde etme amac\u0131yla piyasada hareket etmektedir. BIST Tahminlemede RSI (Relative Strength Index – G\u00f6reli G\u00fc\u00e7 Endeksi) \u0130ndikat\u00f6r\u00fc<\/a> nas\u0131l kullan\u0131ld\u0131 bir \u00f6nceki yaz\u0131m\u0131zda ele alm\u0131\u015ft\u0131k. Tabi ki tek bir teknik indikat\u00f6r her zaman do\u011fru bilgi vermez bu sebeple genellikle birden fazla indikat\u00f6r birlikte de\u011ferlendirilir. Peki perspektifi geni\u015fletirsek ba\u015fka hangi teknik g\u00f6stergeler ve \u00e7evresel \u00f6znitelikler tahminlemede kullan\u0131l\u0131yor birlikte g\u00f6z atal\u0131m.\u00a0<\/span><\/p>\n

\u00d6ncelikle elimizdeki \u00f6znitelikleri d\u00f6rt kategori alt\u0131nda toplayabiliriz. Bunlar RSI, PPO, EMA, CCR gibi teknik indikat\u00f6rleri kapsayan Teknik \u0130ndikat\u00f6rler, D\u00f6viz kurlar\u0131, emitalar, ekonomik performans ve faiz oran\u0131\/para arz\u0131 g\u00f6stergelerini kapsayan Makro-Ekonomi, hisseye ait bilan\u00e7o ve gelir de\u011fi\u015fkenleri, finansal raporlar\u0131 i\u00e7erem \u00c7evresel indikat\u00f6rler ve at\u0131lan tweetlerin sentiment analizi, finansal haber i\u00e7eriklerinin duygu durumlar\u0131, influencerlar\u0131n tavsiyeleri gibi parametreleri kapsayan Sosyal Medya & Haberler \u00fcst ba\u015fl\u0131klar\u0131d\u0131r. Makine \u00f6\u011frenmesi i\u00e7in bu kategoriler genelde g\u00f6r\u00fcnmezlerdir fakat elimizdeki de\u011ferleri kategorilere ay\u0131rmak bizim problemi ele almam\u0131z\u0131 kolayla\u015ft\u0131racakt\u0131r.<\/span><\/p>\n

\"Bist<\/p>\n

Teknik g\u00f6stergeler asl\u0131nda yakla\u015f\u0131k 1400 farkl\u0131 \u00f6zniteli\u011fin tamam\u0131d\u0131r. Fakat makine \u00f6\u011frenmesi ile en s\u0131k kullan\u0131lan teknik g\u00f6stergeler A\u00e7\u0131l\u0131\u015f, Kapan\u0131\u015f, En Y\u00fcksek Fiyat, En D\u00fc\u015f\u00fck Fiyat, Hacim, Basit Hareketli Ortalama (SMA), RSI<\/a>, \u00dcssel Hareket Ortalamas\u0131 (EMA), Hareketli Ortalama Yak\u0131nsama Sapmas\u0131d\u0131r (MACD). Makro ekonomik g\u00f6stergeleri biraz daha detayl\u0131 ele ald\u0131\u011f\u0131m\u0131zda ilk kar\u015f\u0131m\u0131za \u00e7\u0131kan ve en s\u0131k kullan\u0131lan \u00f6znitelik Amerikan Dolar\u0131d\u0131r. Amerikan dolar\u0131\/Japon Yeni (USD\/JPY) aras\u0131ndaki kur oran\u0131 s\u0131kl\u0131kla kullan\u0131lmaktad\u0131r. Dolar\u0131n \u00c7in Yuana oran\u0131 (USD\/CNY),\u00a0 Dolar\u0131n Kanada dolar\u0131na ve Euroya oran\u0131 (USD\/CAD, USD\/EUR) en s\u0131k kullan\u0131lan parametrelerdir. Bununla birlikte ekonomik performans g\u00f6stergelerinden ulusal verimlilik, ticaret ve sekt\u00f6r performans\u0131 , \u00dcFE ve T\u00dcFE, Faiz oranlar\u0131 ve para arz\u0131 bir \u00e7ok yakla\u015f\u0131mda modele dahil edilmi\u015ftir. Makro g\u00f6stergelerden Alt\u0131n ve Petrol emitas\u0131 fiyatlamalar\u0131 yine borsa fiyat modellemede en s\u0131k kullan\u0131lanlardand\u0131r.\u00a0<\/span><\/p>\n

\u00c7evresel \u0130ndikat\u00f6rler yine yakla\u015f\u0131k 150 farkl\u0131 g\u00f6stergeyi kapsamaktad\u0131r fakat en s\u0131k bor\u00e7-toplam varl\u0131klar oran\u0131n\u0131 g\u00f6steren Sermayele\u015fme Oran\u0131, cari oran yani Likide Oran\u0131, varl\u0131k getiri oran\u0131 olan Karl\u0131l\u0131k Oran\u0131 ve toplam varl\u0131k devir oran\u0131 yani Aktivite Oran\u0131d\u0131r. Bu ba\u011flamda, \u015firketin g\u00f6receli karl\u0131l\u0131\u011f\u0131, g\u00f6receli finansman\u0131 ve varl\u0131klar\u0131n\u0131 kullanma yetene\u011fi, hisse ve hisse senedi piyasa tahminleri i\u00e7in mali tablolardan \u00e7\u0131kan en \u00f6nemli de\u011fi\u015fkenler gibi g\u00f6r\u00fcnmektedir.\u00a0<\/span><\/p>\n

Sosyal Medya ve Finansal Haberler ba\u015fl\u0131\u011f\u0131 tahminlemede 2015\u2019ten beri baz\u0131 \u00e7al\u0131\u015fmalarda ele al\u0131nan bir konudur. Huang ve arkada\u015flar\u0131 (2010) \u00e7al\u0131\u015fmalar\u0131nda finansal haber ba\u015fl\u0131klar\u0131n\u0131n analiz etmi\u015flerdir. \u00c7o\u011fu \u00e7al\u0131\u015fma yat\u0131r\u0131mc\u0131 duyarl\u0131l\u0131\u011f\u0131n\u0131 Tayvan elektronik gazeteleri (Huang et al., 2010), Bloomberg (Jin et al., 2013)<\/a> ve FINET (Li et al., 2014) gibi kamuya a\u00e7\u0131k medya kaynaklar\u0131ndan gelen haberleri analiz ederek \u00f6l\u00e7mektedir. Bununla birlikte baz\u0131 bilimsel \u00e7al\u0131\u015fmalar etki sahibi ki\u015filerin g\u00f6r\u00fc\u015flerini do\u011frudan payla\u015ft\u0131\u011f\u0131 sosyal medya i\u00e7eriklerini modele dahil etmi\u015flerdir (Yu, Duan, & Cao, 2013; Wang, Xu, & Zheng, 2018<\/span>)<\/span>.<\/span><\/p>\n

Bu yaz\u0131da literat\u00fcrde s\u0131kl\u0131kla kullan\u0131lan \u00f6znitelikleri kategorik olarak ele ald\u0131k. Data Kapita<\/a>l olarak hem teknik g\u00f6stergelerin bir k\u0131sm\u0131n\u0131 i\u00e7eren backtest sonu\u00e7lar\u0131na hem sosyal medya etki sahibi ki\u015filerin ayl\u0131k s\u0131ralamas\u0131na platform \u00fczerinden ula\u015fabilirsiniz.<\/span><\/p>\n

Burada dikkat \u00e7ekilmesi gereken bir di\u011fer nokta, \u00f6znitelik setlerinin zamanla geni\u015flemesidir. \u00d6zellikle makine \u00f6\u011frenmesi uygulamalar\u0131nda, hangi de\u011fi\u015fkenlerin ger\u00e7ekten anlaml\u0131 bir sinyal ta\u015f\u0131d\u0131\u011f\u0131 konusu h\u00e2l\u00e2 tart\u0131\u015fmal\u0131d\u0131r. \u00d6rne\u011fin, RSI veya MACD gibi klasik teknik indikat\u00f6rlerin k\u0131sa vadede fiyat hareketlerini \u00f6ng\u00f6rmedeki ba\u015far\u0131lar\u0131, \u00e7o\u011fu \u00e7al\u0131\u015fmada s\u0131n\u0131rl\u0131 bulunmu\u015ftur. Buna kar\u015f\u0131n, bu indikat\u00f6rlerin sosyal medya duyarl\u0131l\u0131\u011f\u0131 ya da makroekonomik g\u00f6stergelerle birlikte kullan\u0131lmas\u0131 halinde modellerin do\u011fruluk oranlar\u0131 kayda de\u011fer bi\u00e7imde artmaktad\u0131r. Yani tek ba\u015f\u0131na \u00f6zniteliklerin g\u00fcc\u00fc s\u0131n\u0131rl\u0131yken, \u00e7oklu \u00f6znitelik kombinasyonlar\u0131 daha g\u00fc\u00e7l\u00fc tahmin performans\u0131 sunmaktad\u0131r.<\/p>\n

Bunun yan\u0131nda, \u00f6zniteliklerin farkl\u0131 frekanslarda \u00f6l\u00e7\u00fclmesi de \u00f6nemli bir metodolojik tart\u0131\u015fmad\u0131r. Teknik indikat\u00f6rler genellikle g\u00fcnl\u00fck veya saatlik verilerle hesaplan\u0131rken, bilan\u00e7o g\u00f6stergeleri \u00e7eyrek d\u00f6nemliktir, makroekonomik g\u00f6stergeler ise ayl\u0131k veya y\u0131ll\u0131k frekansta a\u00e7\u0131klanmaktad\u0131r. Bu farkl\u0131 zaman \u00f6l\u00e7eklerini tek bir modele entegre etmek, tahmin performans\u0131n\u0131 do\u011frudan etkiler. Literat\u00fcrde bu uyumsuzlu\u011fu gidermek i\u00e7in \u201cfeature engineering\u201d teknikleri, zaman serisi yeniden \u00f6rnekleme (resampling) ve senkronizasyon y\u00f6ntemleri \u00f6nerilmektedir.<\/p>\n

Sosyal medya ve haber verisinin tahmin g\u00fcc\u00fcne gelince, burada iki boyut \u00f6ne \u00e7\u0131kmaktad\u0131r: hacim ve duygu. Bir hisseye dair sosyal medyada payla\u015f\u0131lan i\u00e7eriklerin hacmindeki art\u0131\u015f, \u00e7o\u011fu zaman fiyat oynakl\u0131\u011f\u0131n\u0131 tetikleyebilir. Buna ek olarak, duygu analizleri sayesinde pozitif, negatif ve n\u00f6tr i\u00e7eriklerin a\u011f\u0131rl\u0131klar\u0131 \u00f6l\u00e7\u00fclerek piyasa alg\u0131s\u0131n\u0131n y\u00f6n\u00fc belirlenebilir. G\u00fcn\u00fcm\u00fczde derin \u00f6\u011frenme tabanl\u0131 do\u011fal dil i\u015fleme y\u00f6ntemleri, klasik duygu analizlerinden \u00e7ok daha y\u00fcksek do\u011fruluk sunmaktad\u0131r. Transformer tabanl\u0131 modeller, \u00f6zellikle finansal dilin ba\u011flamsal yap\u0131s\u0131n\u0131 yakalamakta ba\u015far\u0131l\u0131d\u0131r.<\/p>\n

Son olarak, \u00f6zniteliklerin modellemede nas\u0131l kullan\u0131ld\u0131\u011f\u0131 da kritik bir fakt\u00f6rd\u00fcr. Lojistik regresyon veya basit karar a\u011fa\u00e7lar\u0131 yerine, g\u00fcn\u00fcm\u00fczde ensembllar (XGBoost, Random Forest), derin sinir a\u011flar\u0131 ve hibrit yap\u0131lar daha s\u0131k tercih edilmektedir. Bu y\u00f6ntemler, karma\u015f\u0131k ve y\u00fcksek boyutlu veri setlerini daha iyi i\u015fleyebilmektedir. Ancak daha karma\u015f\u0131k modeller beraberinde \u201ca\u015f\u0131r\u0131 uyum\u201d (overfitting) riskini de getirir. Bu nedenle, d\u00fczenlile\u015ftirme (regularization), \u00e7apraz do\u011frulama ve backtest s\u00fcre\u00e7leri, modelin g\u00fcvenilirli\u011fini sa\u011flamak i\u00e7in kritik \u00f6neme sahiptir.<\/p>\n

\u00d6zetle, teknik, makroekonomik, \u00e7evresel ve sosyal medya temelli \u00f6zniteliklerin hepsi tek ba\u015f\u0131na s\u0131n\u0131rl\u0131 de\u011fer ta\u015f\u0131r. Ancak do\u011fru bi\u00e7imde entegre edildiklerinde, borsa fiyat tahmininde g\u00fc\u00e7l\u00fc bir tahmin kapasitesi sunarlar. Gelecek \u00e7al\u0131\u015fmalarda, \u00f6zellikle \u00e7ok kaynakl\u0131 verinin entegrasyonu ve duygu analizinin rafine edilmesi, tahmin do\u011frulu\u011funu daha da art\u0131racak temel alanlar olarak \u00f6ne \u00e7\u0131kmaktad\u0131r.<\/p>\n

Referanslar<\/strong><\/p>\n

[1] <\/span>Badolia, L. (2016). <\/span>How can i get started investing in the stock market<\/span><\/i>. Educreation Publishing<\/span>.<\/span><\/p>\n

[2] <\/span>Huang, C. J., Liao, J. J., Yang, D. X., Chang, T. Y., & Luo, Y. C. (2010). Realization of a news dissemination agent based on weighted association rules and text mining techniques. <\/span>Expert Systems with Applications<\/span><\/i>, <\/span>37<\/span><\/i>(9), 6409-6413.<\/span><\/p>\n

[3] <\/span>Jin, F., Self, N., Saraf, P., Butler, P., Wang, W., & Ramakrishnan, N. (2013, August). Forex-foreteller: Currency trend modeling using news articles. In <\/span>Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining<\/span><\/i> (pp. 1470-1473).<\/span><\/p>\n

[4] Xiaodong Li, Haoran Xie, Li Chen, Jianping Wang, Xiaotie Deng, News impact on stock price return via sentiment analysis, Knowledge-Based Systems, Volume 69, 2014.<\/span><\/p>\n

[5] <\/span>Yu, Y., Duan, W., & Cao, Q. (2013). The impact of social and conventional media on firm equity value: A sentiment analysis approach. <\/span>Decision support systems<\/span><\/i>, <\/span>55<\/span><\/i>(4), 919-926.<\/span>.<\/span><\/p>\n

[6] <\/span>Wang, Q., Xu, W., & Zheng, H. (2018). Combining the wisdom of crowds and technical analysis for financial market prediction using deep random subspace ensembles. <\/span>Neurocomputing<\/span><\/i>, <\/span>299<\/span><\/i>, 51-61.<\/span>.<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"

Ortalama bir insan\u0131n borsaya olan ilgisi ge\u00e7ti\u011fimiz y\u0131llarda \u00fcssel olarak artm\u0131\u015ft\u0131r [1]. Yani herg\u00fcn milyarlarca dolar de\u011ferinde varl\u0131k kar elde etme amac\u0131yla piyasada hareket etmektedir. BIST Tahminlemede RSI (Relative Strength Index – G\u00f6reli G\u00fc\u00e7 Endeksi) \u0130ndikat\u00f6r\u00fc nas\u0131l kullan\u0131ld\u0131 bir \u00f6nceki yaz\u0131m\u0131zda ele alm\u0131\u015ft\u0131k. Tabi ki tek bir teknik indikat\u00f6r her zaman do\u011fru bilgi vermez bu<\/p>\n","protected":false},"author":13,"featured_media":5100,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[39,6,41],"tags":[480,482],"class_list":["post-5099","post","type-post","status-publish","format-standard","has-post-thumbnail","category-analiz-teknikleri","category-finansal-veri-okuryazarligi","category-veri-turleri-ve-kavramlar","tag-bist","tag-teknik-indikatorler"],"better_featured_image":{"id":5100,"alt_text":"Bist \u00d6znitelik \u0130nceleme","caption":"","description":"","media_type":"image","media_details":{"width":626,"height":418,"file":"2023\/12\/Bist-teknik-indikatorler.jpg","filesize":132209,"sizes":{"medium":{"file":"Bist-teknik-indikatorler-300x200.jpg","width":300,"height":200,"mime-type":"image\/jpeg","filesize":19661,"source_url":"https:\/\/datakapital.com\/blog\/wp-content\/uploads\/2023\/12\/Bist-teknik-indikatorler-300x200.jpg"},"thumbnail":{"file":"Bist-teknik-indikatorler-150x150.jpg","width":150,"height":150,"mime-type":"image\/jpeg","filesize":8310,"source_url":"https:\/\/datakapital.com\/blog\/wp-content\/uploads\/2023\/12\/Bist-teknik-indikatorler-150x150.jpg"},"bunyad-small":{"file":"Bist-teknik-indikatorler-150x100.jpg","width":150,"height":100,"mime-type":"image\/jpeg","filesize":6067,"source_url":"https:\/\/datakapital.com\/blog\/wp-content\/uploads\/2023\/12\/Bist-teknik-indikatorler-150x100.jpg"},"bunyad-medium":{"file":"Bist-teknik-indikatorler-450x300.jpg","width":450,"height":300,"mime-type":"image\/jpeg","filesize":38650,"source_url":"https:\/\/datakapital.com\/blog\/wp-content\/uploads\/2023\/12\/Bist-teknik-indikatorler-450x300.jpg"}},"image_meta":{"aperture":"0","credit":"","camera":"","caption":"","created_timestamp":"0","copyright":"","focal_length":"0","iso":"0","shutter_speed":"0","title":"","orientation":"0","keywords":[]}},"post":5099,"source_url":"https:\/\/datakapital.com\/blog\/wp-content\/uploads\/2023\/12\/Bist-teknik-indikatorler.jpg"},"amp_enabled":true,"_links":{"self":[{"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/posts\/5099","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/users\/13"}],"replies":[{"embeddable":true,"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/comments?post=5099"}],"version-history":[{"count":3,"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/posts\/5099\/revisions"}],"predecessor-version":[{"id":5542,"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/posts\/5099\/revisions\/5542"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/media\/5100"}],"wp:attachment":[{"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/media?parent=5099"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/categories?post=5099"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/datakapital.com\/blog\/wp-json\/wp\/v2\/tags?post=5099"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}