-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathindex.html
More file actions
708 lines (448 loc) · 24.7 KB
/
index.html
File metadata and controls
708 lines (448 loc) · 24.7 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
<!DOCTYPE html>
<html>
<head>
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta name="description" content="Biomedical informatics research fellow with interests in machine learning and clinical analytics.">
<meta name="keywords" content="bernauer, biomedical, informatics, data mining, machine learning, statistics, R, SQL, Python">
<meta name="google-site-verification" content="fF5Xy64LacKRhQoTmTmVPb8WT59AfwVvPYXEJJalrBU" />
<title>Michael L. Bernauer</title>
<link rel="shortcut icon" type="image/x-icon" href="/img/portrait_crop.png"/>
<link type="text/css" rel="stylesheet" media="all" href="/css/reset.css"/>
<link type="text/css" rel="stylesheet" media="all" href="/css/default.css"/>
<script type="text/javascript" src="http://ajax.googleapis.com/ajax/libs/jquery/1.4.3/jquery.min.js"></script>
<script type="text/javascript" src="js/smoothscroll.js"></script>
<link href='http://fonts.googleapis.com/css?family=Oswald' rel='stylesheet' type='text/css'>
</head>
<body>
<div id="header">
<div class="transparent">
<div class="nav">
<ul>
<li><a href="#home">Home</a></li>
<li><a href="#about">About</a></li>
<li><a href="#projects">Projects</a></li>
<li><a href="assets/bernauer_curriculum_vitae.pdf">Curriculum Vitae</a></li>
<li><a href="pages/downloads">Downloads</a></li>
<li><a href="pages/slides/">Slides</a></li>
<li><a href="#contact">Contact</a></li>
</ul>
</div>
</div>
</div>
<!--HOME-->
<div id="home">
<div class="beginning">
<h1>Michael L. Bernauer</h1>
</div>
</div>
<!--ABOUT-->
<div id="about">
<div class="beginning">
<h2>about me</h2>
</div>
<div class="wrap-one">
<div class="doublegulp">
<p class="about">Biomedical informatics research fellow with interests in machine learning and clinical analytics. </p>
</div>
</div>
</div>
<div id="projects">
<div class="beginning">
<h2>projects</h2>
</div>
<div class="page-content">
<div class="wrapper">
<div class ="wrap">
<div class="biggulp">
<img src="./assets/python-data-summarization.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20160320-data-summarization-in-python.html"> Data Summarization in Python</a></h3>
<h4>Mar 20, 2016</h4>
<p>In this example I show how python can be used for various data summarization and data visualization tasks using a few essential libraries such as <b>pandas</b>, <b>ggplot</b>, <b>seaborn</b>, and <b>matplotlib</b>. In the following notebook we will use these libraries to compute some summary statistics on data from the Centers for Medicare and Medicaid Services (CMS). Specifically, we will look at top drugs by total spend, total number of claims and total number of beneficiaries. We will the visualizing our results using <strong>seaborn</strong>, <strong>ggplot</strong>, and <strong>matplotlib</strong>.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="assets/R.jpg" width=450 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20160319-naive-bayes-information-retrieval.html"> Information Retrieval using Naive Bayes</a></h3>
<h4>Mar 19, 2016</h4>
<p>Naive Bayes Classifiers have been used historically for document classification (i.e. spam detection). They belong to a class of probablistic classifiers and are considered 'naive' due to their assumption of independence between features. Naive Bayes works by taking a set of features and computing the joint conditional probability of these features over as set of labels, P(features|label). It then uses Bayes theorem to return the lable with the largest probabiltiy conditioned on the feature set, P(lable|features).</p>
<p>In this notebook we will implement a Bayes classifier using R and use it to return documents given a query, it can also be used to determine author attribution. For example, given an unlabeled text, return a list of possible authors. Additionally, it may be possible to use this method to determine which journal a particular abstract should be submitted.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./assets/ari-histogram.svg" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20160308-automated-readability-index-R.html"> Automated Readability Index</a></h3>
<h4>Mar 8, 2016</h4>
<p>The Automated Readability Index (ARI)is a measure that is used to estimate the reading
level required to understand a text. Using average word length, and average
sentence length, the ARI returns the estimated grade level of the text. In this notebook
we will implement the ARI function and use it to estimate the grade level of several scientific
abstracts published in PubMed.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./assets/pubmed-auths.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20160304-pubmed-author-clustering-R.html"> PubMed Authors: Hierarchical Clustering</a></h3>
<h4>Mar 4, 2016</h4>
<p>In this notebook we take a look at <a href="http://www.pubmed.gov">PubMed</a> authors and how they can
be clustered into groups using Term Frequency-Inverse Document Frequency weighted terms extracted
from article titles. In particular, we will focus on clustering authors from the University of New Mexico
College of Pharmacy and the School of Medicine.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="assets/python.png" width=450 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20160131-document-retrieval-sklearn.html"> Machine Learning - Python: Document Similarity</a></h3>
<h4>Jan 31, 2016</h4>
<p>In this notebook we will implement a k-Nearest Neighbors model from sklearn to return similar
scientific articles using TF-IDF weights as features. The articles used to create the corpus
in this notebook are downloaded from <a href="http://www.ncbi.nlm.nih.gov/pubmed/?term=%22university+of+new+mexico%22%5BAD%5D+AND+(%22pharmacy%22%5BAD%5D+OR+%22medicine%22%5BAD%5D)">pubmed.gov</a>
and restricted to articles submitted by the University of New Mexico College of Pharmacy and
the University of New Mexico School of Medicine.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="assets/R.jpg" width=450 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20160117-mlr-w3-polynomial-regression.html"> Machine Learning - R: Polynomial Regression</a></h3>
<h4>Jan 16, 2016</h4>
<p>This post describes the principles of polynomial regression and looks at the
bias-variance trade-off for different models of varying complexity.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="assets/R.jpg" width=450 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20160110-multiple-linear-regression-2.html"> Machine Learning - R: Simple Multivariable Linear Regression</a></h3>
<h4>Jan 10, 2016</h4>
<p>This post describes an R implementation of simple, multivariable linear regression
in which parameters optimizing Residual Sum of Squares (RSS) are determined using
gradient descent.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./assets/cms_claims.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./pages/cms/"> Analysis of Medicare Part D Drug Claims Data</a></h3>
<h4>Jan 2, 2016</h4>
<p>For this project, I was interested in studying drug claims and their associated cost to the Medicare system. This was was done by aggregating the number of claims, beneficiaries and total cost for each of the records listed in the PUF. I then took a subset of the top 200 drugs according to number of claims submitted to Medicare in order to find the most commonly prescribed drugs.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="assets/python.png" width=450 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20151230-mlr-w1-simple-regression-python.html"> Machine Learning - Python: Simple Linear Regression</a></h3>
<h4>Dec 30, 2015</h4>
<p>This post describes a python implementation of simple, one-variable linear regression
in which parameters optimizing Residual Sum of Squares (RSS) are determined by
the closed form method.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="assets/R.jpg" width=450 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20151230-mlr-w1-simple-regression-R.html"> Machine Learning - R: Simple Linear Regression</a></h3>
<h4>Dec 30, 2015</h4>
<p>This post describes an R implementation of simple, one-variable linear regression
in which parameters optimizing Residual Sum of Squares (RSS) are determined by
the closed form method.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./assets/20151115_daily_drug.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="http://dailydrug.ddns.net/"> Top 200 Daily drug email list</a></h3>
<h4>Nov 15, 2015</h4>
<p>Over the last couple of months I've been working with the <a href="https://www.cms.gov/">CMS</a> Medicare Part D dataset in which I've been able to exctact the top 200 most
commonly prescribed medications in New Mexico. Using this list and a copy of the
<a href="http://www.drugbank.ca/">DataBank</a> database I was able to create a mailing list
that will provide a "drug of the day" email providing drug information for one
of the top 200 drugs. To sign-up, visit <a href="http://dailydrug.ddns.net">http://dailydrug.ddns.net</a> and add your email to the list.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./img/nis-race-small.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./pages/NIS/"> Statewide vaccination rates: National Immunization Survey</a></h3>
<h4>Aug 13, 2015</h4>
<p>Immunizations are important. They help keep communities safe by preventing the spread
of disease. However, to be most effective, a certain percentage of the population must
be vaccinated in order to achieve herd immunity. This figure looks at data from the
CDC National Immunization Survey to show vaccination rates across the United States.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150705_hypothesis_testing.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150705_hypothesis_testing.html"> Population inference and hypothesis testing</a></h3>
<h4>Jul 5, 2015</h4>
<p>How do we use statistics to make inferences from a sample to
a larger population? How confident are we that these statistics
accurately represent these population parameters and how can we use
statistics to test whether or not our observations are different
from hypothesized values. These are some of the questions that
will be addressed in this article.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150630_python_cell_detection.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150630_python_cell_detection.html"> Statistical approach to cell detection using Python</a></h3>
<h4>Jun 30, 2015</h4>
<p>This is a project conducted in Python in which I was interested in
counting cells in an image.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150630_cop_pubmed_rec_small.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150630_cop_pubmed_rec_large.png"> University of New Mexico College of Pharmacy Publication Record</a></h3>
<h4>Jun 30, 2015</h4>
<p>This figure shows all publications indexed in PubMed beginning from
1988 through December of 2014. Of course this only includes articles
that have been indexed in PubMed, so these numbers may not reflect the
total number of articles published for that given time.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150608_medicare_part_d.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150608_medicare_part_d.html"> Analysis of Medicare Part D 2012 data</a></h3>
<h4>Jun 8, 2015</h4>
<p>This example presents an exploratory analysis of 2012 Medicare Part D data.
The data set contains over 21 million rows and was collected from ProPublica.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150321_decision_analysis.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150321_pharmacoeconomics_decision_analysis.html"> Pharmacoeconomics: Sensitivity Analysis</a></h3>
<h4>Mar 14, 2015</h4>
<p>In the example we are interested in computing the incremental cost per life year and the
incremental cost per QALY to aid in decision making between two treatment options;
Treatment A, and Treatment B. We then conduct a sensitivity analysis to observe how these
measures change in response to variations in input parameters; probability of experiencing a
particular event and the costs associated with the event.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150314_logistic_regression.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150314_logistic_regression.html"> Logistic Regression</a></h3>
<h4>Mar 14, 2015</h4>
<p>In this exercise we create a binomial logistic regression model to predict whether someone will default on their
credit card based on their current credit card balance and income. Once the model is built, we use it to plot the
decision boundary for defaulting. Additionally, we see that the overall accuracy of the model is 90%.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150314_ISLR_linear_regression.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150314_ISLR_linear_regression.html"> ISLR: Linear Regression</a></h3>
<h4>Mar 14, 2015</h4>
<p>In this exercise we work the chapter 3 of <a href="http://www-bcf.usc.edu/%7Egareth/ISL/">Introduction to Statistical Learning</a> by Trevor Hastie and colleagues to explore
linear regression and how it can be used to predict the median house value of
neighborhoods based on the average number of rooms per dwelling, the average age of the house as well as the percentage of houses that qualify as low socioeconomic status.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150309_visualize_this_patterns_over_time.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150309_visualize_this_patterns_over_time.html"> Visualize This: Visualizing patterns over time</a></h3>
<h4>Mar 9, 2015</h4>
<p>In this exercise we work with data provided by Nathan Yau, author of <a href="http://flowingdata.com/">www.flowingdata.com</a>
to visualize trends in United States unemployment data from 1948-2010 using LOESS.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150309_treemaps_with_ggplot2.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150309_treemaps_with_ggplot2.html"> Tree-maps with ggplot2</a></h3>
<h4>Mar 9, 2015</h4>
<p>In this exercise we learn how to create Treemaps with ggplot2 using the <a href="https://github.com/wilkox/treemapify">Treeplotify</a> package.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150308_visualize_this_patterns_over_time.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150308_visualize_this_patterns_over_time.html"> Visualize This: Visualizing patterns over time</a></h3>
<h4>Mar 8, 2015</h4>
<p>In this exercise we work with data provided by Nathan Yau, author of <a href="http://flowingdata.com/">www.flowingdata.com</a>
to visualize the number of hot dogs and buns (HDBs) eaten by contestants of the Nathans hot dot eating competition from 1980-2010.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150228_coursera_ml_neural_networks.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150228_coursera_ml_neural_networks.html"> Machine Learning: Multi-class classification with neural networks</a></h3>
<h4>Feb 28, 2015</h4>
<p>In this post we explore the use of artificial neural networks as multi-class classifiers.
In the example provided we will use previously learned parameters to identify handwritten
digits using the feed forward algorithm. Neural networks are preferred to the logistic
regression methods mentioned in previous posts for these sorts of tasks because they are
capable of forming highly complex decision boundaries without the need to train
computationally expensive high order models.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150225_coursera_ml_multiclass_classification_logistic_regression.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150225_coursera_ml_multiclass_classification_logistic_regression.html"> Machine Learning: Multi-class classification with logistic regression</a></h3>
<h4>Feb 27, 2015</h4>
<p>This exercise implements a one-vs-all linear regression classifier to recognize handwritten
digits using gradient descent. Regularized gradient descent was introduced in the post <a href="http://mlbernauer.bitbucket.org/R/20150222_coursera_ml_regularized_logistic_regression.html">Regularized Logistic Regression</a>. Once trained, the model was able to recognize handwritten digits with an accuracy of 83.8%. It is anticipated that the accuracy
could be improved by allowing more complete convergence of the cost function. </p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150226_stat_ada_graphical_summaries.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150226_stat_ada_graphical_summaries.html"> Advanced Data Analysis: Graphical Summaries in R</a></h3>
<h4>Feb 26, 2015</h4>
<p>This post walks you through a few of the most common plots to graphically summarize data
using <em>ggplot2</em>.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150224_stat_ada_introduction_to_R.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150224_stat_ada_introduction_to_R.html"> Advanced Data Analysis: Introduction to R</a></h3>
<h4>Feb 24, 2015</h4>
<p>R is an open-source statistical computing language used for a wide range of tasks from data analysis and visualization to machine learning. In this exercise we work through the first chapter of Advanced Data Analysis; a supplemental text to the course STAT 527 - Advanced Data Analysis as taught by <a href="http://statacumen.com/">Erik Erhardt, PhD</a> at the University of New Mexico. For those of you that are interested in data science I highly recommend taking his course if possible or at least reading through the material posted on his website.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150222_coursera_ml_regularized_logistic_regression.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150222_coursera_ml_regularized_logistic_regression.html"> Machine Learning: Regularized Logistic Regression</a></h3>
<h4>Feb 21, 2015</h4>
<p>In this exercise we will implement regularized logistic regression to predict
whether microchips from a fabrication plant passes quality assurance (QA).
During QA, each microchip goes through various tests to ensure it is functioning
correctly.</p>
<p>Suppose you are the product manager of the factory and you have the test results
for some microchips on two different tests. From these two tests, you would like
to determine whether the microchips should be accepted or rejected. To help
make the decision you have a data set of test results on past microchips, from which
can build a logistic regression model.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150221_coursera_ml_logistic_regression.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150221_coursera_ml_logistic_regression.html"> Machine Learning: Logistic Regression</a></h3>
<h4>Feb 21, 2015</h4>
<p>In this exercise we will build a logistic regression model to
predict whether a student gets admitted into a university.
Suppose that you are the administrator of a university department and
you want to determine each applicants chance of admission based on their
results on two exams. We have historical data from previous applicants
that will be used as a training set for logistic regression. For each training
example, we have the applicants scores on two exams and the admissions
decision.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./R/20150221_coursera_ml_linear_regression.png" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./R/20150221_coursera_ml_linear_regression.html"> Machine Learning: Linear Regression</a></h3>
<h4>Feb 21, 2015</h4>
<p>In this exercise we will be implementing linear regression with one variable using the
gradient descent algorithm. Given data collected by a food truck we will attempt
to predict profit using city population as a predictor.</p>
</div>
</div>
<div class ="wrap">
<div class="biggulp">
<img src="./img/data_cleansing_small.jpg" width=550 align='middle'/>
</div>
<div class="gulp">
<h3><a href="./img/data_cleansing.jpg"> Data Cleansing: FDA Adverse Event Reporting System</a></h3>
<h4>Dec 23, 2014</h4>
<p>Many data sets are inherently dirty, they contain misspellings, missing data, erroneous values, etc. This is particularly true for data sets created by people typing information into forms or spreadsheets manually. The FDA Adverse Event Reporting System is just that type of data set. Spend just a few minutes browsing the data and you will see that it is rife with errors. For example, consider the drug metformin. In the current database I am working on there are 84,459 different mentions of metformin. By grouping and counting these mentions we can see the top 10 most frequent representations in the table below. More than half of all mentions of metformin are represented as METFORMIN HCL whereas about 24% of the mentions are represented simply as METFORMIN.</p>
</div>
</div>
</div>
</div>
</div>
<!--CONTACT-->
<div id="contact">
<div class="beginning">
<h2>contact</h2>
</div>
<div class="wrap-one">
<div class="social">
<p class= "about">If you want to get in touch, email is probably best. But you have options.</p>
<a href="mailto:mlbernauer@gmail.com" class="mailicon" title="mail"></a>
<a href="http://twitter.com/mlbernauer" class="twittericon" title="twitter"></a>
<a href="http://www.facebook.com/mbernauer" class="facebookicon" title="facebook"></a>
<a href="https://github.com/mlbernauer" class="githubicon" title="github"></a>
<a href="https://www.flickr.com/photos/mlbernauer/" class="flickricon" title="flickr"></a>
</div>
</div>
</div>
<!--FOOTER-->
<div id="footer"></div>
</body>
</html>