Manjari

Advanced Data Analysis Project
Linking galaxies to their progenitors based on galaxy morphology

1. Introduction
2. Data
3. Methods
4. Exploratory Data Analysis
   4.1 Mass rank at z=1 and z=2
   4.2 Changes in mass rank from z=1 to z=2
   4.3 Galaxy morphology at z=1
5. Predicting mass Rank at an Earlier Epoch
   5.1 Predicting mass rank at z=2 given data at z=1
6 Linking galaxies at z = 2 to galaxies at z = 1
   6.1 Checking performance of linking method with simulated data
       6.1.1 Two-sample paired t-test
       6.1.2 Compare k-nearest neighbor with remaining galaxies
   6.2 Checking performance of linking method with real data
7 Summary and Conclusions
Appendices

---------------------------------------------------------------------

Kevin

Covariance-based sample selection under
heterogenous data for autism risk gene
detection

1 Introduction
2 Data and model background
  2.1 Modeling approach
  2.2 Connections
3 Elementary analysis
  3.1 Specification of covariance hypothesis test
  3.2 Application to BrainSpan
4 Methods (*)
  4.1 Stepdown procedure: multiple testing with dependence
  4.2 Computational extension for stepdown procedure
  4.3 Largest partial clique: selecting partitions based on testing
      results
5 Simulation study
6 Application on BrainSpan study
  6.1 Partition selection
  6.2 Gene network and detected risk genes
  6.3 Investigation on detected risk genes
7 Conclusion and discussions

Appendices

---------------------------------------------------------------------

Justin

Valid post-selection inference for Segmentation Methods with
application to copy number variation data

1 Introduction
  1.1 Notation
  1.2 Changepoint algorithms
  1.3 Changepoint inference
      1.3.1 Existing work
      1.3.2 Setup for changepoint inference framework
2 Post selection inference for changepoint problems
  2.1 General case
  2.2 Modifications to the general case
  2.3 Practicalities
3 Array CGH data changepoint inference
4 Simulations
  4.1 Synthetic simulations
  4.2 Pseudo-real data application
5 Conclusions

Appendices (*)

---------------------------------------------------------------------

Fuchen

A framework of change detection for dynamic
networks

1 Introduction
2 Background information
  2.1 Stochastic block model (SBM)
  2.2 Global spectral clustering for dynamic networks
3 Change point detection algorithms
  3.1 Related works
  3.2 Problem setting
  3.3 Eigen test
  3.4 Likelihood ratio test
  3.5 Multiple change points
  3.6 Theoretical results for likelihood ratio test
      3.6.1 Single change point
      3.6.2 Multiple change points3.6.2 Multiple change points
4 Trend detection and outlier detection
  4.1 Smooth trend detection
  4.2 Outlier detection
5 Application in real data
  5.1 resting state fMRI data
  5.2 Gene co-expression networks
6 Next steps

Appendices

---------------------------------------------------------------------

Boyan

Modeling the population¡¯s perception
of security over time
in Democratic Republic of Congo

1 Introduction
2 The survey and data
  2.1 Sampling process and studying unit
  2.2 Survey content and pre-process
  2.3 EDA
      2.3.1 Sample size
      2.3.2 Demographic information
      2.3.3 Questions of interest
3 Methods (*)
  3.1 Generating index
  3.2 A model for change over time
  3.3 Assessing relationships
  3.4 Understanding item responses
4 Results
  4.1 Generating indexes (**)
      4.1.1 Overall dimensionality
      4.1.2 Forming indexes
  4.2 Modeling the change overtime
      4.2.1 Security level about personal experience (long)
      4.2.2 Security level about freedom of speech
      4.2.3 Ethnic relations
  4.3 Assessing relationships
      4.3.1 Personal experience security (long index) VS ethnic
      relations
      4.3.2 Freedom of speech security VS ethnic relations
  4.4 Understanding items and their answers
5 Discussion

Appendices

---------------------------------------------------------------------

Ben

Heterogeneous Sub-Population Proportion
Estimations of Immune Cells

1 Introduction
  1.1 Statistical Motivation (*)
2 Dataset
  2.1 Protein Expression Collection (*)
  2.2 Missingness
3 Methods
  3.1 Batch (Mouse Specific) Correction
  3.2 Estimation of Cell Classes via Random Forest
  3.3 Estimating Proportions of Cell Sub-Populations
4 Results
  4.1 Batch Corrections
  4.2 Estimating Proportions
  4.3 Use in Collaborator¡¯s Analysis (*)
5 Discussion and Future Work

(no appendices!)

---------------------------------------------------------------------