圖像梯度增強(qiáng)

背景 (Background)

Purpose of analysis:

分析目的：

Understand the factors driving student success so that Open University can allocate resources to improve student success

了解推動(dòng)學(xué)生成功的因素，以便開(kāi)放大學(xué)可以分配資源來(lái)提高學(xué)生的成功

Description of dataset:

數(shù)據(jù)集說(shuō)明：

The Open University Learning Analytics Dataset is a publicly available dataset containing data about courses, students and their interactions with VLE for seven selected courses (modules).

開(kāi)放大學(xué)學(xué)習(xí)分析數(shù)據(jù)集是一個(gè)公共可用的數(shù)據(jù)集，其中包含有關(guān)課程，學(xué)生及其與七個(gè)所選課程(模塊)的VLE交互的數(shù)據(jù)。

As the unique identifiers across all the data tables were student ID, module and course description, data was aggregated at this level. The following variables were used in the analysis:

由于所有數(shù)據(jù)表中的唯一標(biāo)識(shí)符是學(xué)生ID，模塊和課程說(shuō)明，因此在此級(jí)別匯總了數(shù)據(jù)。 分析中使用了以下變量：

Variable NameVariable TypeVariable NameVariable Typeid_student unique identifier/primary keyimd band categorical code_modulecategoricalage band categorical code_presentation categorical num of previous attemptsnumericalgender categorical studied credits categorical region categorical disability categorical highest education categorical final result numerical sum weighted score numerical average module length numerical average submission duration numerical average proportion content accessed numerical average date registration numerical trimmed assessment type categorical

變量名稱變量類型變量名稱變量類型 id_student唯一標(biāo)識(shí)符/主keyimd頻帶類別code_module類別ageage類別類別code_presentation類別先前嘗試的次數(shù)numericalgender類別classical所學(xué)習(xí)的學(xué)分類別區(qū)域categorical殘疾categorical最高學(xué)歷類別最終平均結(jié)果數(shù)值平均值數(shù)值總和訪問(wèn)的內(nèi)容數(shù)值平均日期注冊(cè)數(shù)值修剪評(píng)估類型類別

方法 (Methodology)

數(shù)據(jù)轉(zhuǎn)換 (Data transformation)

• Combine datasets based on unique identifier (student ID, course and code_description)
  根據(jù)唯一標(biāo)識(shí)符(學(xué)生ID，課程和code_description)組合數(shù)據(jù)集
• Aggregate variables at unique identifier level
  在唯一標(biāo)識(shí)符級(jí)別匯總變量
• Update nominal variable types from character to factor
  將名義變量類型從字符更新為因子

Predictors:

預(yù)測(cè)變量：

• code module
  代碼模塊
• code presentation (changed to 1 = B and 2 = J),
  代碼表示(更改為1 = B和2 = J)，
• gender, region, highest education, IMD band, Age band, number of previous attempts, studied credits,
  性別，地區(qū)，最高學(xué)歷，IMD頻段，年齡段，以前的嘗試次數(shù)，學(xué)分，
• average submission duration (averaged as a single code presentation can have multiple assessments with varying submission durations),
  平均提交時(shí)間(平均一次代碼演示可以進(jìn)行多次評(píng)估，且提交時(shí)間有所不同)，
• average module length,
  平均模塊長(zhǎng)度
• average proportion content accessed (based on sum of clicks on “ou” content and resources/quiz/glossary divided by total sum of clicks per code description),
  平均訪問(wèn)的內(nèi)容比例(基于對(duì)“ ou”內(nèi)容和資源/測(cè)驗(yàn)/詞匯的點(diǎn)擊總和除以每個(gè)代碼描述的總點(diǎn)擊總和)，
• average date registration and
  平均日期注冊(cè)和
• trimmed assessment type (as a single code description can have multiple assessments and multiple of the same type of assessment. It is important to determine whether the types of assessments per course and description drive student success)
  修整的評(píng)估類型(因?yàn)閱蝹€(gè)代碼描述可以具有多個(gè)評(píng)估，也可以具有多個(gè)相同類型的評(píng)估。確定每門課程和描述的評(píng)估類型是否能促進(jìn)學(xué)生成功很重要)

Excluded variables: Student ID and sum weighted score

排除的變量：學(xué)生ID和加權(quán)總和

Reason for exclusion: Student ID– identifying information, sum weighted score — correlated to final result

排除原因：學(xué)生證-識(shí)別信息，總和加權(quán)分-與最終結(jié)果相關(guān)

Define “success” target variable

定義“成功”目標(biāo)變量

• Response variable: Success (“Pass” or “Distinction” = “Yes”, “Fail” or “Withdrawn” = “No”
  響應(yīng)變量：成功(“通過(guò)”或“不同” =“是”，“失敗”或“撤回” =“否”

Reason for not using “final result” as success factor: There is a disproportionate of “Pass” records within the dataset thus reducing the model accuracy at predicting “Fail”, “Withdrawn” and “Distinction”. As such, the response variable was binarised.

不使用“最終結(jié)果”作為成功因素的原因：數(shù)據(jù)集中的“通過(guò)”記錄不成比例，因此降低了預(yù)測(cè)“失敗”，“撤回”和“區(qū)別”時(shí)的模型準(zhǔn)確性。 這樣，將響應(yīng)變量二值化。

• Check for nulls, missing values, and outliers for each variable that will be input into the model

  檢查將輸入到模型中的每個(gè)變量的空值，缺失值和離群值

Why do we check for missing data? — If a large proportion of data is missing/null, sample is not representative enough to provide accurate results. This is not the case when the missing/null data is missing for a valid reason.

為什么我們要檢查丟失的數(shù)據(jù)？ —如果大部分?jǐn)?shù)據(jù)丟失/為空，則樣本的代表性不足以提供準(zhǔn)確的結(jié)果。 當(dāng)丟失/空數(shù)據(jù)由于正當(dāng)原因而丟失時(shí)，情況并非如此。

資料設(shè)定 (Data set-up)

? Split data into training & test sets for nominal and binary predictors to test for model accuracy

?將數(shù)據(jù)分為用于名義和二進(jìn)制預(yù)測(cè)變量的訓(xùn)練和測(cè)試集，以測(cè)試模型的準(zhǔn)確性

Why do we split the data into training and test sets — The model is trained on the training set. Model accuracy is tested on the test set to determine how well the model is good at prediction success against non-success on data it has not “seen”.

我們?yōu)槭裁磳?shù)據(jù)分為訓(xùn)練集和測(cè)試集-在訓(xùn)練集上訓(xùn)練模型。 在測(cè)試集上測(cè)試模型準(zhǔn)確性，以確定模型針對(duì)未“看到”的數(shù)據(jù)不成功的預(yù)測(cè)成功率。

資料建模 (Data modelling)

• Analytical method: Gradient boosting machine (GBM)

  分析方法：梯度提升機(jī)(GBM)

• Alternate methods could have been used but model accuracy was good enough — Distributed Random Forest (DRF) and Generalized Linear Model (GLMNET)
  可以使用其他方法，但是模型的精度足夠好-分布式隨機(jī)森林(DRF)和廣義線性模型(GLMNET)

• Check for model accuracy using confusion matrix (i.e. proportion that were predicted correctly) and area under curve (how well did we do compared to random chance).

  使用混淆矩陣(即正確預(yù)測(cè)的比例)和曲線下面積(與隨機(jī)機(jī)會(huì)相比，我們做得如何)檢查模型準(zhǔn)確性 。

輸出量 (Output)

• Top predictors (by variable importance score)
  最佳預(yù)測(cè)變量(按重要性變量計(jì))

數(shù)據(jù)ETL (Data ETL)

All data exploration, transformation, and loading of final dataset was done in Alteryx to avoid writing code and test its functionality for basic data transformation steps that would typically be carried out in R such as joins, mutate and group by.

所有數(shù)據(jù)探索，轉(zhuǎn)換和最終數(shù)據(jù)集的加載均在Alteryx中完成，以避免編寫(xiě)代碼并測(cè)試其功能，以進(jìn)行通常在R中執(zhí)行的基本數(shù)據(jù)轉(zhuǎn)換步驟，例如連接，變異和分組。

First, I joined three datasets — assessments.csv, studentAssessments.csv and courses.csv by ID assessments (primary key).

首先，我通過(guò)ID評(píng)估(主鍵)加入了三個(gè)數(shù)據(jù)集-Assessments.csv，studentAssessments.csv和courses.csv。

Next, I engineered two features: weighted_score and submission_duration. Dates are usually meaningless unless they are transformed into useful features. Course description only had two categories, which I converted to 1 and 2 for ease of reference.

接下來(lái)，我設(shè)計(jì)了兩個(gè)功能：weighted_score和Submitting_duration。 日期通常是沒(méi)有意義的，除非將其轉(zhuǎn)換為有用的功能。 課程描述只有兩類，為了便于參考，我將其轉(zhuǎn)換為1和2。

The next step was to add in data on student interactions with the virtual learning platform (VLE). Some feature engineering was done for ease of analysis such as creating a new variable called activity_type_sum that reduces the number of categories in activity type into two broad categories — content access and browsing. The reason for doing this is that granular categories only result in more features and reduce the number of observations per category. The number of clicks were summed by the activity type feature. Proportion of activity out of total activity that is browsing related and content access related was also calculated. This is a good way to create a feature that is relative to another feature and scaled by total activity thus ensuring that all students are represented on a similar scale by their activity type.

下一步是添加有關(guān)學(xué)生與虛擬學(xué)習(xí)平臺(tái)(VLE)互動(dòng)的數(shù)據(jù)。 為了簡(jiǎn)化分析，進(jìn)行了一些功能設(shè)計(jì)，例如創(chuàng)建一個(gè)名為activity_type_sum的新變量，該變量將活動(dòng)類型中的類別數(shù)量減少為兩大類-內(nèi)容訪問(wèn)和瀏覽。 這樣做的原因是，細(xì)化類別只會(huì)帶來(lái)更多功能，并減少每個(gè)類別的觀測(cè)次數(shù)。 點(diǎn)擊次數(shù)是按活動(dòng)類型功能求和的。 還計(jì)算了與瀏覽相關(guān)和與內(nèi)容訪問(wèn)相關(guān)的總活動(dòng)中活動(dòng)的比例。 這是創(chuàng)建相對(duì)于另一個(gè)功能并按總活動(dòng)進(jìn)行縮放的功能的好方法，從而確保所有學(xué)生按其活動(dòng)類型以相似的比例表示。

Block 1 was joined to Block 2 using student_id, code_module and code_presentation as the primary key. The resulting output is shown below.

使用student_id，code_module和code_presentation作為主鍵，將塊1加入到塊2。 結(jié)果輸出如下所示。

The above output — Block 3 — was joined with student registration data using student_id, code_module, and code_presentation to bring across the data_registered field.

上面的輸出(第3塊)通過(guò)使用student_id，code_module和code_presentation與學(xué)生注冊(cè)數(shù)據(jù)結(jié)合在一起，以顯示data_registered字段。

The date_unregistered field was ignored as it had a lot of missing values. Moreover, students with empty unregistered field cells have withdrawal as the value for their final_result. This variable is our target/response variable. So, the date_unregistered field appears to be a proxy measure for final_result and as such it makes sense to exclude this variable from our analysis.

date_unregistered字段被忽略，因?yàn)樗鄙俸芏嘀怠?此外，未注冊(cè)字段單元格為空的學(xué)生可以將他們的final_result值取回 。 此變量是我們的目標(biāo)/響應(yīng)變量。 因此， date_unregistered字段似乎是final_result的代理度量，因此從我們的分析中排除此變量是有意義的。

As shown above, for a given id_student, code_module, and code_presentation, the module_presentation length, proportion_content and date_registration is repeated. As we want to have unique records, we can aggregate the data as follows:

如上所示，對(duì)于給定的id_student，code_module和code_presentation，重復(fù)module_presentation的長(zhǎng)度，比例內(nèi)容和date_registration。 由于我們希望擁有唯一的記錄，因此可以按以下方式匯總數(shù)據(jù)：

• Summarise weighted score using total sum
  使用總和總結(jié)加權(quán)分?jǐn)?shù)
• Average submission duration
  平均提交時(shí)間
• Average module presentation (you can also use other aggregates such as minimum, maximum, and median)
  平均模塊展示(您也可以使用其他匯總，例如最小值，最大值和中位數(shù))
• Average of proportion_content_access
  ratio_content_access的平均值
• Average of date_registration
  date_registration的平均值

Data is now at student_id, code_module, code_presentation and assessment_type level; however, the target variable — final_result — is at student_id, code_module and code_presentation level. Hence, this data will need to be further aggregated.

數(shù)據(jù)現(xiàn)在位于student_id，code_module，code_presentation和評(píng)估類型級(jí)別； 但是，目標(biāo)變量final_result處于student_id，code_module和code_presentation級(jí)別。 因此，該數(shù)據(jù)將需要進(jìn)一步匯總。

Let’s look at student info first. A unique record here is id_student, code_module, code_presentation. So, we will need to go back a step and summarise a student_id, code_module and code_presentation to represent all assessments taken by an individual. We will still use the previous summary formulas.

我們先來(lái)看一下學(xué)生信息。 這里的唯一記錄是id_student，code_module，code_presentation。 因此，我們將需要退后一步，總結(jié)一個(gè)student_id，code_module和code_presentation來(lái)代表一個(gè)人進(jìn)行的所有評(píng)估。 我們?nèi)詫⑹褂靡郧暗膮R總公式。

By doing this we have 8 unique assessment types that a student can take for a given code module and code presentation. Assessment types are not repeated (trimmed only) so if at student took 3 TMAs this is not reflected as shown below.

通過(guò)這樣做，我們可以為學(xué)生提供針對(duì)給定代碼模塊和代碼表示形式的8種獨(dú)特的評(píng)估類型。 評(píng)估類型不會(huì)重復(fù)(僅修整)，因此，如果學(xué)生參加了3次TMA，則不會(huì)如下所示。

A variable could be created to count the number of assessments per assessment type but it would contain lots of missing values as not all assessments have all three types of assessments. Now, we are ready to join to the student info data with output shown below.

可以創(chuàng)建一個(gè)變量來(lái)計(jì)算每種評(píng)估類型的評(píng)估數(shù)量，但是它會(huì)包含很多缺失值，因?yàn)椴⒎撬性u(píng)估都具有這三種評(píng)估類型。 現(xiàn)在，我們準(zhǔn)備加入學(xué)生信息數(shù)據(jù)，輸出如下所示。

Now, we have 18 columns. We have been told that a presentation may differ if presented in February vs. October. We will assume that it does not differ year on year (i.e. 2013B is same as 2014B). As such, we will recode code_presentation into 1 for B and 2 for J as a binary variable.

現(xiàn)在，我們有18列。 我們被告知，如果在2月和10月之間進(jìn)行演示，則演示可能會(huì)有所不同。 我們將假定它沒(méi)有同比差異(即2013B與2014B相同)。 這樣，我們將將code_presentation重新編碼為B的1和J的2作為二進(jìn)制變量。

The final output is shown below.

最終輸出如下所示。

It is finally time for some data exploration.

現(xiàn)在是時(shí)候進(jìn)行一些數(shù)據(jù)探索了。

探索性數(shù)據(jù)分析 (Exploratory Data Analysis)

Categorical variables can be represented with bar charts where the y-axis is the frequency of the occurence of a given category. For example, in the chart below we can see that the most frequently taken code module is FFF followed by BBB. There are seven unique code modules with no missing values.

類別變量可以用條形圖表示，其中y軸是給定類別的出現(xiàn)頻率。 例如，在下面的圖表中，我們可以看到最常用的代碼模塊是FFF，其次是BBB。 有七個(gè)唯一的代碼模塊，沒(méi)有缺失值。

Data can also be summarised numerically using a five-point summary for continuous variables and using mode for categorical variables as shown below.

數(shù)據(jù)也可以使用五點(diǎn)匯總(用于連續(xù)變量)和使用模式(用于分類變量)以數(shù)字方式匯總，如下所示。

Insights we can make from the summary below is that the more common student is a Scottish male student presenting without a disability with an Imd_band between 20 to 40% with a typical Pass as their final result.

我們可以從下面的摘要中得出的見(jiàn)解是，最普通的學(xué)生是蘇格蘭的一名男生，表現(xiàn)出無(wú)障礙，Imd_band在20％到40％之間，最終成績(jī)?yōu)檎！?

Now we can move towards modelling the dataset.

現(xiàn)在我們可以對(duì)數(shù)據(jù)集建模。

機(jī)器學(xué)習(xí)模型 (Machine Learning Model)

We have been asked to assist Open University in better understanding student success. ? We will assume that student success is measured via the final result where pass and distinction are indicators of “success” and withdrawn and fail are indicators of “non-success”. ? For the independent variables, we will use all variables from the previoustable except for weighted_score. The reason for this is because weighted score determines the final result for a given student. As such, it is highly correlated (multicollinear) to final result and as such will be excluded. ? Student ID is identifying information and as such will not be used as a predictor.

我們被要求協(xié)助開(kāi)放大學(xué)更好地了解學(xué)生的成功。 ?我們將假設(shè)學(xué)生的成功是通過(guò)最終結(jié)果來(lái)衡量的，合格和區(qū)別是“成功”的指標(biāo)，退縮和失敗是“不成功”的指標(biāo)。 ?對(duì)于自變量，我們將使用上一張表中的所有變量(weighted_score除外)。 這樣做的原因是因?yàn)榧訖?quán)分?jǐn)?shù)決定了給定學(xué)生的最終結(jié)果。 因此，它與最終結(jié)果高度相關(guān)(多重共線性)，因此將被排除。 ?學(xué)生證是識(shí)別信息，因此不會(huì)用作預(yù)測(cè)指標(biāo)。

GBM (Gradient Boosted Model) was used as a model of choice. This type of model creates a series of weak learners (shallow trees) where each new tree tries to improve on the error rate of the previous tree. The final tree is one with the lowest error rate. It is an ensemble machine learning method as several trees are created to provide the final results. However, unlike in randomForest, these trees are created in a series rather than in parallel. Furthermore, these trees are not independent and are depenent on the previous tree’s error rate where the following three will try harder to improve prediction for the more difficult cases. This is controlled by a parameter called hte learning rate.

使用GBM(梯度增強(qiáng)模型)作為選擇模型。 這種類型的模型會(huì)創(chuàng)建一系列弱學(xué)習(xí)者(淺樹(shù))，其中的每棵新樹(shù)都試圖提高前一棵樹(shù)的錯(cuò)誤率。 最終的樹(shù)是錯(cuò)誤率最低的樹(shù)。 這是一種集成的機(jī)器學(xué)習(xí)方法，因?yàn)閯?chuàng)建了幾棵樹(shù)以提供最終結(jié)果。 但是，與randomForest不同，這些樹(shù)是按系列而不是并行創(chuàng)建的。 此外，這些樹(shù)不是獨(dú)立的，而是取決于前一棵樹(shù)的錯(cuò)誤率，在后一棵樹(shù)上，后三棵將更努力地提高對(duì)更困難情況的預(yù)測(cè)。 這是由稱為學(xué)習(xí)率的參數(shù)控制的。

The model was run with 500 rounds (500 trees) with minimum and maximum depths of 4 for the tree. Typically, it is not good to have very deep trees as this can lead to overfitting where the algorithm tries to explain every observation in the dataset as it increases the depth of the tree leading to leaves containing a very small number of observations that fit the given rule.

該模型以500輪(500棵樹(shù))運(yùn)行，最小和最大深度為4。 通常，擁有非常深的樹(shù)不是很好，因?yàn)檫@會(huì)導(dǎo)致過(guò)度擬合，因?yàn)樗惴〞?huì)嘗試解釋數(shù)據(jù)集中的每個(gè)觀察值，因?yàn)樗鼤?huì)增加樹(shù)的深度，從而導(dǎo)致葉子中包含非常少的符合給定觀察值的觀察值規(guī)則。

We can see from the above output that the model has a RMSE (root-mean-squared error) value of 0.55 which is quite high. It is particularly bad at predicting Distinction and Fail, which may be due to the imbalance in the dataset where we know from our exploratory data analysis that Pass is the most common final result.

從上面的輸出中我們可以看到，該模型的RMSE(均方根誤差)值為0.55，這非常高。 在預(yù)測(cè)Distinction和Fail時(shí)特別糟糕，這可能是由于數(shù)據(jù)集的不平衡所致，從探索性數(shù)據(jù)分析中我們知道Pass是最常見(jiàn)的最終結(jié)果。

To counteract this imbalance issue, the target variable was redefined as “success” (distinction and pass) and “failure” (fail and withdrawn). It is common to combine categories to deal with imbalanced datasets. Other ways are to undersample (i.e. reduce the number of instances for the most frequent class) or oversample (i.e. create artificial observations for the non-frequent classes).

為了解決此不平衡問(wèn)題，將目標(biāo)變量重新定義為“成功”(區(qū)分并通過(guò))和“失敗”(失敗并撤回)。 合并類別以處理不平衡的數(shù)據(jù)集是很常見(jiàn)的。 其他方法是欠采樣(即減少最頻繁分類的實(shí)例數(shù)量)或過(guò)采樣(即為非頻繁分類創(chuàng)建人工觀察)。

The model was re-run with the following output. Here we can see that the mean per-class error has dropped significantly. The Area Under the Curve (AUC) is another accuracy metric that tells you how well the model is at classifying cases correctly (i.e. maximising the true positive rate (TPR)). The higher the AUC, the more accurate the model. As the AUC is measured between 0 and 1, an AUC of 0.87 is pretty good.

使用以下輸出重新運(yùn)行模型。 在這里，我們可以看到平均每類錯(cuò)誤已大大降低。 曲線下面積(AUC)是另一種精度度量標(biāo)準(zhǔn)，它告訴您模型在正確分類案例(即最大化真實(shí)正利率(TPR))方面的表現(xiàn)。 AUC越高，模型越準(zhǔn)確。 由于測(cè)得的AUC在0到1之間，因此0.87的AUC相當(dāng)不錯(cuò)。

Another metric that is commonly used in classification problems is the F1 score which is the harmonic mean of precision and recall. Both metrics aim to maximise the TPR while minimising either the false negative rate (recall) or false positive rate (precision). A true positive is when a success or failure is classified correctly. A false negative is when a success is labelled as a failure. A false positive is when a failure is labelled as a success. For the F1 score to be high, both precision and recall need to be high.

分類問(wèn)題中通常使用的另一個(gè)度量標(biāo)準(zhǔn)是F1分?jǐn)?shù)，它是精確度和查全率的諧波平均值。 兩種指標(biāo)均旨在最大程度地提高TPR，同時(shí)最大程度地降低誤報(bào)率(召回率)或誤報(bào)率(精度)。 正確的判斷是對(duì)成功或失敗進(jìn)行了正確分類。 假否定是將成功標(biāo)記為失敗。 誤報(bào)是指將失敗標(biāo)記為成功。 為了使F1分?jǐn)?shù)高，準(zhǔn)確性和召回率都必須高。

Confusion matrix indicates an overall error rate of 17.11% which is mainly driven by how good the model is at classifying successes. The model is not so good at classifying failures with an error rate of 39.07%. Again this may be due to the data being overrepresented by “Passes”. Thus, the results should be treated with caution and model re-run with a more balanced dataset.

混淆矩陣表明總體錯(cuò)誤率為17.11％，這主要是由模型對(duì)成功分類的良好程度驅(qū)動(dòng)的。 該模型不能很好地對(duì)故障進(jìn)行分類，錯(cuò)誤率為39.07％。 同樣，這可能是由于數(shù)據(jù)被“通行證”過(guò)多所代表。 因此，應(yīng)謹(jǐn)慎對(duì)待結(jié)果，并使用更平衡的數(shù)據(jù)集重新運(yùn)行模型。

Now, let’s look at the top predictors of success or failure by looking at the variable importance list.

現(xiàn)在，讓我們通過(guò)查看變量重要性列表來(lái)查看成功或失敗的主要預(yù)測(cè)因素。

• The top 3 variables are code module, trimmed_assessment_type and average submission duration.
  前三個(gè)變量是代碼模塊，trimmed_assessment_type和平均提交時(shí)間。
• The bottom 3 variables at predicting whether or not a successful outcome will be reached for a given student are disability status, average date for registration and gender.
  預(yù)測(cè)給定學(xué)生是否達(dá)到成功結(jié)果的最后3個(gè)變量是殘疾狀況，平均注冊(cè)日期和性別。

• Note: As code module and code presentation are part of the unique identifier they should have been excluded from analysis. However, as presentations in February and October may be different for some courses, both variables were kept in the model. It is possible that excluding these variables may increase accuracy or make other variables more “important”.

  注意：由于代碼模塊和代碼表示是唯一標(biāo)識(shí)符的一部分，因此應(yīng)將它們從分析中排除。 但是，由于某些課程在2月和10月的演示文稿可能有所不同，因此兩個(gè)變量都保留在模型中。 排除這些變量可能會(huì)提高準(zhǔn)確性或使其他變量更“重要”。

Now, let’s visualise information on the top predictors to better understand the model. The stacked bar plot below shows the proportion of records by course module and final_result. We can deduce that students are more likely to be successful in completing AAA, EEE and GGG courses over other courses.

現(xiàn)在，讓我們可視化頂級(jí)預(yù)測(cè)變量上的信息，以更好地理解模型。 下面的堆疊條形圖按課程模塊和final_result顯示了記錄的比例。 我們可以推斷出，與其他課程相比，學(xué)生更有可能成功完成AAA，EEE和GGG課程。

• From the above table, we can see that there is a 100% success rate if an exam is the only assessment for a given course and presentation.
  從上表中可以看出，如果考試是給定課程和演示的唯一評(píng)估，則成功率為100％。
• If only computer marked assessments make up the course component, there is a very high failure/withdrawn rate. It would be interesting to investigate why having CMA as part of a presentation assessment leads to a decrease in success rate.
  如果僅由計(jì)算機(jī)標(biāo)記的評(píng)估構(gòu)成課程組成部分，則故障/退出率非常高。 調(diào)查為什么將CMA作為演示評(píng)估的一部分會(huì)導(dǎo)致成功率降低，這將是很有趣的。

The histograms above show the average submission duration by success and failure.

上面的直方圖顯示了成功和失敗的平均提交時(shí)間。

It appears that when students are successful, they are more likely to submit their assignment within 10 days (+/-) of the assessment submission date.

看來(lái)，當(dāng)學(xué)生成功時(shí)，他們更有可能在評(píng)估提交日期的10天內(nèi)(+/-)提交作業(yè)。

結(jié)語(yǔ) (Wrapping up)

Machine learning was used to quickly identify top contributors to student success.

機(jī)器學(xué)習(xí)用于快速確定學(xué)生成功的主要貢獻(xiàn)者。

Recommendations for model improvement include:

有關(guān)模型改進(jìn)的建議包括：

• Working with a balanced dataset
  使用平衡的數(shù)據(jù)集
• Including proxy measures for resource allocation within the dataset
  包括數(shù)據(jù)集中資源分配的代理指標(biāo)

• Count of the number of assessments by type per course and presentation as a

  按課程類型和演示形式按類型進(jìn)行的評(píng)估數(shù)量計(jì)數(shù)

  feature

  特征

• Remove categorical variables that are associated with each other (i.e. use of chi-squared test of independence)
  刪除彼此關(guān)聯(lián)的分類變量(即，使用卡方獨(dú)立性檢驗(yàn))

Hopefully, you now have a better understanding of utilising GBM for a classification problem, the pitfalls of a classification problem (i.e. imbalanced dataset) and the use of various accuracy metrics.

希望您現(xiàn)在對(duì)使用GBM解決分類問(wèn)題，分類問(wèn)題的陷阱(即數(shù)據(jù)集不平衡)以及使用各種準(zhǔn)確性指標(biāo)有了更好的了解。

The reference to all R code is provided in my git repository: https://github.com/shedoesdatascience/openlearning

在我的git存儲(chǔ)庫(kù)中提供了對(duì)所有R代碼的引用： https : //github.com/shedoesdatascience/openlearning

翻譯自: https://towardsdatascience.com/using-gradient-boosting-machines-for-classification-in-r-b22b2f8ec1f1

圖像梯度增強(qiáng)

總結(jié)

以上是生活随笔為你收集整理的图像梯度增强_使用梯度增强机在R中进行分类的全部?jī)?nèi)容，希望文章能夠幫你解決所遇到的問(wèn)題。

如果覺(jué)得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。