SMHS LinearModeling MachineLearning - Revision history

Pineaumi: /* References */

2016-05-23T13:58:32Z

‎References

Pineaumi: /* SMHS Linear Modeling - Machine Learning Algorithms */

2016-05-23T13:55:10Z

‎SMHS Linear Modeling - Machine Learning Algorithms

Imoubara at 12:57, 23 May 2016

2016-05-23T12:57:26Z

Pineaumi: /* Example 1: Simulation (subject, day, treatment, observation) */

2016-05-19T12:49:50Z

‎Example 1: Simulation (subject, day, treatment, observation)

Pineaumi: /* Example 1: Simulation (subject, day, treatment, observation) */

2016-05-19T12:48:44Z

‎Example 1: Simulation (subject, day, treatment, observation)

Pineaumi: /* References */

2016-05-19T12:48:18Z

‎References

Pineaumi: /* SMHS Linear Modeling - Machine Learning Algorithms */

2016-05-19T12:45:09Z

‎SMHS Linear Modeling - Machine Learning Algorithms

Pineaumi: /* Example 2: Genotype-phenotype */

2016-05-19T12:43:34Z

‎Example 2: Genotype-phenotype

Pineaumi: /* SMHS Linear Modeling - Machine Learning Algorithms */

2016-05-19T12:40:26Z

‎SMHS Linear Modeling - Machine Learning Algorithms

Pineaumi: /* SMHS Linear Modeling - Machine Learning Algorithms */

2016-05-18T16:09:10Z

‎SMHS Linear Modeling - Machine Learning Algorithms

← Older revision		Revision as of 13:58, 23 May 2016
Line 627:		Line 627:

	Winter, B. (2013). Linear models and linear mixed effects models in R with linguistic applications. arXiv:1308.5499. [http://arxiv.org/pdf/1308.5499.pdf]		Winter, B. (2013). Linear models and linear mixed effects models in R with linguistic applications. arXiv:1308.5499. [http://arxiv.org/pdf/1308.5499.pdf]
−
−
−	~~....~~

@@ Line 35: / Line 35: @@
   predicted.values <-  predict(lin.mod, <b><u>newdata=test</u></b>
-Data Modeling/Training
+<b>Data Modeling/Training</b>
 <b>Logistic Regression:</b>

@@ Line 4: / Line 4: @@
 <b>Questions:</b>
-<li>How can we tie human intuition and computer-generated results to obtain reliable, effective, and efficient decision-support system (that facilitates, forecasting)?</li>
+*How can we tie human intuition and computer-generated results to obtain reliable, effective, and efficient decision-support system (that facilitates, forecasting)?
-<li>Niels Born – “It is difficult to make predictions, especially about the future”</li>
+*Niels Born – “It is difficult to make predictions, especially about the future”
-<li>Can we unsupervisely classify the data?</li>
+*Can we unsupervisely classify the data?
 <b>Prediction</b>
 For most of the machine learning algorithms (including first-order linear regression), we:
-<li>ﬁrst generate the model using training data, and then</li>
+*ﬁrst generate the model using training data, and then
-<li>predict values for test/new data. </li>
+*predict values for test/new data.
 Predictions are made using the R <b>predict</b> function. (type <b>?predict.name</b>), where <b>name</b> is the function-name corresponding to the algorithm. The ﬁrst argument of predict often represents the variable storing the model, and the second argument is a matrix or data frame of test data that the model needs to be applied to. Calling predict can be done in 2 ways: type <b>predict</b> or type of <b>predict.name.</b>

@@ Line 91: / Line 91: @@
 <i>Obs ~ Treatment + Day + Subject(Treatment)+ Day*Subject(Treatment)+ ε.</i>
-This model is accounts for:
+This model is accounts for:<BR>
 Response = Obs
-Fixed effects:
+Fixed effects:<BR>
+Treatment (fixed)<BR>
-Treatment (fixed)
+Day (fixed)<BR>
 Treatment*Day interaction
-Random Effects:
+Random Effects:<BR>
+Subject nested within Treatment (random)<BR>
 Day crossed with "Subject within Treatment" (random)

@@ Line 89: / Line 89: @@
 ===Example 1: Simulation (subject, day, treatment, observation)===
-<blockquote><i>Obs ~ Treatment + Day + Subject(Treatment)+ Day*Subject(Treatment)+ ε.</i></blockquote>
+<i>Obs ~ Treatment + Day + Subject(Treatment)+ Day*Subject(Treatment)+ ε.</i>
 This model is accounts for:

← Older revision		Revision as of 12:48, 19 May 2016
Line 637:		Line 637:

	....		....
−

@@ Line 14: / Line 14: @@
 <li>predict values for test/new data. </li>
-Predictions are made using the R <b>predict</b> function. (type <b>?predict.name</b>), where <b>name</b> is the function-name corresponding to the algorithm. The ﬁrst argument of predict often represents the variable storing the model and the second argument is a matrix or data frame of test data that the model needs to be applied to. Calling predict can be done in 2 ways: type <b>predict</b> or type of <b>predict.name.</b>
+Predictions are made using the R <b>predict</b> function. (type <b>?predict.name</b>), where <b>name</b> is the function-name corresponding to the algorithm. The ﬁrst argument of predict often represents the variable storing the model, and the second argument is a matrix or data frame of test data that the model needs to be applied to. Calling predict can be done in 2 ways: type <b>predict</b> or type of <b>predict.name.</b>
 <b>Example:</b>

@@ Line 548: / Line 548: @@
   data$\$$y <- simulate(model.lmer)    # param bootstrap data set
   model.lmer.1 <- lmer(unlist(data$\$$y) ~ g + (1|t), data= data)
- # we need to turn the list of simulated values (y) into an atomic vector with unlist()
+   # we need to turn the list of simulated values (y) into an atomic vector with unlist()
- # both data frames and models objects (e.g.,  produced by lm()) are lists
+   # both data frames and models objects (e.g.,  produced by lm()) are lists
- #  http://adv-r.had.co.nz/Data-structures.html
+   #  http://adv-r.had.co.nz/Data-structures.html
   gdiff[i] <- fixef(model.lmer.1)[2]   # keep only the est diff
+  }
   quantile(gdiff, c(0.025, 0.975))
- # print 95% CI of the coefficient/effect-size of genotype(g)
+   # print 95% CI of the coefficient/effect-size of genotype(g)
   # print model summary
@@ Line 580: / Line 580: @@
   # Using gdt, we can obtain a parametric bootstrap-t interval
+  # Hypothesis testing: here we need to simulate under <b><u>H0: g = 0</u></b>
   # To find the t statistic and the F statistic
   #  t statistic is in the @coefs table
@@ Line 607: / Line 606: @@
   # obs F = 2.0684
   anova(model.lmer)
- mean(gdt0[,2]$\$$sim_2$\$$residuals >= 2.0684)
+   mean(gdt0[,2]$\$$sim_2$\$$residuals >= 2.0684)
   # p-value = 0.01785714
 <b>Baseball data</b>
 http://wiki.socr.umich.edu/index.php/SOCR_Data_MLB_HeightsWeights
   data <- read.table('E:\\Ivo.dir\\Research\\UMichigan\\Education_Teaching_Curricula\\2015_2016\\HS_853_Fall_2015\\Modules_docx\\01a_data.txt',as.is=T, header=T)

@@ Line 85: / Line 85: @@
   summary(svm_model)
-Appendix
+==Appendix==
-Example 1: Simulation (subject, day, treatment, observation)
+===Example 1: Simulation (subject, day, treatment, observation)===
 <blockquote><i>Obs ~ Treatment + Day + Subject(Treatment)+ Day*Subject(Treatment)+ ε.</i></blockquote>
@@ Line 358: / Line 358: @@
 </center>
-Example 2: Genotype-phenotype
+===Example 2: Genotype-phenotype===
 Save this data file as a tab-separated TXT file:
@@ Line 622: / Line 622: @@
   main = "MLB Team Weight Distributions")
-<b>References</b>
+===References===
 Bates, D.M., Maechler, M., & Bolker, B.  (2012). lme4:  Linear mixed-effects models using S4 classes. R package version 0.999999-0.

← Older revision		Revision as of 16:09, 18 May 2016
Line 637:		Line 637:

	Winter, B. (2013). Linear models and linear mixed effects models in R with linguistic applications. arXiv:1308.5499. [http://arxiv.org/pdf/1308.5499.pdf]		Winter, B. (2013). Linear models and linear mixed effects models in R with linguistic applications. arXiv:1308.5499. [http://arxiv.org/pdf/1308.5499.pdf]
−
−
−
−
−
−
−
−
−
−
−
−