匿名使用者
匿名使用者 發問時間: 社會與文化語言 · 1 0 年前

中英夾雜翻譯..請諸位大大協助一下,非常感謝

如題..非常感激~~ ^^

所有的方法在G function的experiment下,皆可使A獲得良好的學習,而得到非常approximation其函數曲線(function curve)的結果。

由於在all experiments中所比較的3 learning algorithms中,每種algorithm在執行一次iteration的時間不同,若以固定的學習epoch來做比較將較不具客觀性。Thus, 本研究於all experiments中,將給予每種learning algorithm在20 min.的execution time內去做學習,同時並設定一夠大的學習epoch(4,000次),以做為stop condition. 而在所比較的algorithms中,能execution幾次(iteration)就讓其盡量去做學習,待所設定的execution time停止後,才去就每種algorithm於最終所得到the number of the best Y hidden nodes下之結果做比較,而得到如Fig. x所示的learning curves比較情形。

已更新項目:

Sorry..未詳細說明

是中翻"英"啦...多謝多謝!!

4 個解答

評分
  • 最佳解答

    所有的方法在G function的experiment下,皆可使A獲得良好的學習,而得到非常approximation其函數曲線(function curve)的結果。

    由於在all experiments中所比較的3 learning algorithms中,每種algorithm在執行一次iteration的時間不同,若以固定的學習epoch來做比較將較不具客觀性。Thus, 本研究於all experiments中,將給予每種learning algorithm在20 min.的execution time內去做學習,同時並設定一夠大的學習epoch(4,000次),以做為stop condition. 而在所比較的algorithms中,能execution幾次(iteration)就讓其盡量去做學習,待所設定的execution time停止後,才去就每種algorithm於最終所得到the number of the best Y hidden nodes下之結果做比較,而得到如Fig. x所示的learning curves比較情形。

    All methods are under experiment of G function, can all make A get good study , and it is extraordinary to receive the result of its function curve of approximation (function curve ). Because in 3 learning algorithms compared in all experiments, each algorithm carry out one time of iteration different , is it is it have objectivity with regular study epoch to make to come. Thus, this research will offer each kind of learning algorithm in 20 min in all experiments. Execution time in do it study,at the same time more establish(4,000 times) one study epoch not very big,regard making as by stop condition. Among compared algorithms, can execution several (iteration ) let it is it is it study to do to try one's best, after execution time established stops, just go to compare in the results under the number of the best Y hidden nodes got finally on each kind of algorithm, and get like Fig. X learning curves comparative situation shown.

    希望我有幫助到你

  • 1 0 年前

    All means' experiment under is at G function, all can make A acquire good study, but get very approximation it function curve (function curve)Result.

    Because is comparative 3 learning algorithms in in all experiments, each kind algorithm is executing once iterations' time different, if as fixed study epoch to makes very will more not objectification. Thus, this research in all experiments in, will grant each kind learning algorithm at 20 min. Execution time in goes to make study, meanwhile set up a big study epoch enough and (4, 000 times), as make for stop condition. But is at in comparative algorithms, can execution several times (iteration)Let it try the best go to make study, treat set up execution time stop behind, just go to then each kind algorithm in get the number of the best Y hidden nodes under result make very finally, but get be like Fig. x show learning curves compare situation.

    =========================================================

    大概是這個樣子啦!

    參考資料: 自翻+字典
  • 1 0 年前

    咍囉

    All methods can make the AN acquire a good learning, and get very under the experiment of G function the result of the approximation its function curve(function curve).

    Because in the all experiments compare of grow algorithm the time dissimilarity been carrying out an iteration each time in the algorithms of the 3 learning, if with the learning epoch fixed do relatively will compare not to have objectivity.Thus, this research will give to grow learning algorithm each time in all experiments at 20 mins.Of do a learning inside execution time, also set one enough big learning epoch(4,000 times) in the meantime to be used as stop condition. And in the algorithms compared, the ability execution lets it do a learning as far as possible several times(iteration), needing the execution time set stop after, just go to grow algorithm each time in final gain the hidden nodes of the number of the best Y it do a comparison as a result, and get as Fig. The comparison situation of learning curves that the x shows.

    還有問題加我及時

    suzycsptw

    要註明你是要翻譯的人喔

    ^^"

    參考資料: hoping I can help you!!
  • 1 0 年前

    所有的方法在【G作用】的【實驗】下,皆可使A獲得良好的學習,而得到非常【略計】其函數曲線【作用曲線】的結果。

    由於在【所有實驗】中所比較的【3種學習算法】中,每種【算法】在執行一次【疊代】的時間不同,若以固定的學習【世紀】來做比較將較不具客觀性。【因而】, 本研究於【所有實驗】中,將給予每種【學習算法】在【20分鐘】.的【執行時間】內去做學習,同時並設定一夠大的學習【世紀】(4,000次),以做為【停止狀態】. 而在所比較的【算法】中,能【施行】幾次【(疊代)】就讓其盡量去做學習,待所設定的【執行時間】停止後,才去就每種【算法】於最終所得到【最佳的Y 暗藏的結的數量】下之結果做比較,而得到如【圖x】所示的【經驗曲線】比較情形。

    參考資料: 自己
還有問題?馬上發問,尋求解答。