Occupant Activity Profiles from Smart Home Sensor Event Streams

class: center, middle, inverse, title-slide

# Occupant Activity Profiles from Smart Home Sensor Event Streams
### Garrick Aden-Buie<br><i>University of South Florida</i>
### <span style="font-size: 60%">Ali Yalcin, Ph.D.<br>Carla VandeWeerd, Ph.D.</span>
### INFORMS Annual Meeting<br>Oct 25, 2017

---

exclude: true
background-image: url('images/ruth-individual.JPG')
background-size: cover

???

Early in 2012, just after her 80th birthday my grandmother was diagnosed with Alzheimer’s.

Same time: grandfather hospitalized for heart failure. Also COPD.

They had just re-retired to Florida, 1200 miles from Chicago where my parents were living.

My mom left a full-time position and for a while would fly back and forth.

My grandparents were highly independent, very much wanted to live on their own, but it was very difficult for my mom to monitor them from a distance.

As Alz progressed, tougher to know if they ate, showered, left the house, slept normally, etc.

Two highlights from their story

1. Not uncommon (aging stats and discussion)...
2. Highlighting the need and value for passive technologies that link caregivers and seniors

---
## Older adults live longer but have more chronic health issues

![](images/older_adults.png)

- Life expectancy +3 months every year since 1840

- 75% older adults report being in good to excellent health

- 70% report chronic health conditions

.footnote[
Wouters 2017; Christensen 2009; Older Americans 2012
]

---
## Growing older population need assistance leading to care gaps

![](images/older_adults.png)

- US health care system faces shortfalls in primary care physicians and paid caregivers

- 50% need or receive help with routine activities

- .highlight[Baby boomers:] 45M adults 65+ as of 2015

.footnote[
Bodenheimer 2009; Scommegna 2016; US Census 2015
]

---
layout: true
## US Population Age Distribution
---

![](images/pop-dist.gif)

<span style="font-size: 0.5em">
<https://www.weforum.org/agenda/2016/03/167-years-of-the-us-age-demographic-in-one-chart>
</span>

---

![](images/pop-dist-together.png)

<span style="font-size: 0.5em">
<https://www.weforum.org/agenda/2016/03/167-years-of-the-us-age-demographic-in-one-chart>
</span>

---
layout: false
class: inverse
background-image: url('images/senior-at-home.jpg')
background-size: cover

## Aging in Place

<div style="background-color: rgba(0,0,0,0.3); padding:15px;"><strong>The ability to live in one's own home and community <em>safely, independently, and comfortably</em>.</strong></div>

???
96% of older adults live independently in homes/apartments

75% strongly agree: want to stay in current residence for as long as possible

Need help: majority of the help is family. Changes in family dynamics: more older adults without living spouses, or adult children within 10 miles.

"Self-care of chronic conditions will therefore be an important component of the aging process, while family members will feel increased pressure to make up for gaps in care availability"

---
class: doctor-senior

## Aging in Place

96% of older adults live independently in homes/apartments

75% want to stay in current residence for as long as possible

**Benefits**

- Maintain independence, autonomy and social connections
- Reduce health care costs
- Increase quality of life, limit effects of aging

**Trade-Offs**

- Inconsistent monitoring
- Delayed health assessments

.footnote[
CDC Healthy Places; Mather2015
]

???

reduced health care costs: avoid expensive institutional care

limit aging effects: cognitive decline and depression

inconsistent monitoring: not seeing physician regularly, fear of institutionalization, etc.

---
layout: false
class: mom-daughter

## Lifestyle Reassurance Systems

- Use in-home, passive wireless motion, contact and appliance sensors

- Allow older adults to .highlight[age in place]

- Monitor health and daily activities of older adult

- Connect caregivers and health- and elder-care providers

???

In order to enable long-term management of chronic conditions while aging in place, there is a need for technological solutions in the home that facilitate self-care activities, connect older adults with primary care physicians and elder care service providers as well as informal family caregivers.

---
## In-home sensor systems face significant challenges

1. Difficult to capture and label realistic, nature patterns of daily living

2. Passive sensor systems need to be minimally invasive

3. Low-cost and easy to deploy and maintain at large scales

---
layout: true
class: large
## Research Objective
---

Can we effectively use passively-recorded, 
unlabelled sensor data to characterize<br>
lifestyle activity patterns to enable <br>
detection of changes in routine?

---

Can we effectively use passively-recorded, 
.highlight[unlabelled sensor data] to characterize<br>
.highlight[lifestyle activity patterns] to enable<br>detection of changes in routine?

---
layout: false
class: large
## Overview

1. HomeSense

2. Activity Profiles

3. Occupant Identification using<br> Activity Profiles

---
class: bottom, right
background-image: url("images/homesense.jpg")
background-size: cover

# HomeSense

---
layout: false
## Study Participants

**Health BOOST**, _University of South Florida_

- Recruitment
    - 55+ resident of The Villages
    - Living alone in pet-free home with internet access
    - Do not exhibit signs of cognitive impairment

- 14 total participants (7 ongoing)
    - 5,450 total days of observation

- 12 have completed at least 6 months

---
## Participants considered for comparison of Activity Profiles

- 7 Participants

- Completed at least 196 days (~7 months):
    - Excluding vacations, maintenance days
    - At "home" for most of the day
    - Avoiding guests and visitors

- Similar floorplan and sensor layout

---
exclude: true
## Sensors

.pull-left[.center[
##### Motion
![](images/sensors/motion.png)
##### Contact
![](images/sensors/contact2.png)
]]

.pull-right[.center[
##### Power and Water
![](images/sensors/power.png)
##### System
![](images/sensors/system.png)
]]

---
layout: true
## Home and Sensor Layout
---

exclude: true
background-image: url("images/floorplans/blank.png")
background-position: 50% 45%
background-size: 100%

---

exclude: true
background-image: url("images/floorplans/simple-legend.png")
background-position: 50% 45%
background-size: 100%

---

layout: false
background-image: url("images/floorplans/simple-exp-covered.png")
background-position: 50% 45%
background-size: 100%
## Sensor Layout & Coverage

---
## Daily Activity Heatmap

<img src="gab-informs2017_files/figure-html/plot-calendar-1.png" width="720" />

---
layout: false
class: inverse, center, middle
# Activity Profiles
---

# Activity Profile Process

- Gather sensor event sequences

- Group into *day-length* event sequences

- Build *activity profiles* from day-event sequences

---
layout: true
## Sensor Event Sequences
---

exclude: true

<table>
 <thead>
  <tr>
   <th style="text-align:left;"> Timestamp </th>
   <th style="text-align:left;"> Sensor </th>
   <th style="text-align:right;"> State </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:01:44 </td>
   <td style="text-align:left;"> InLivingRoom </td>
   <td style="text-align:right;"> 1 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:05:28 </td>
   <td style="text-align:left;"> InLivingRoom </td>
   <td style="text-align:right;"> 1 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:26:25 </td>
   <td style="text-align:left;"> InKitchen </td>
   <td style="text-align:right;"> 1 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:28:34 </td>
   <td style="text-align:left;"> OpenedGarageDoor </td>
   <td style="text-align:right;"> 1 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:28:35 </td>
   <td style="text-align:left;"> InLivingRoom </td>
   <td style="text-align:right;"> 1 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:28:58 </td>
   <td style="text-align:left;"> OpenedGarageDoor </td>
   <td style="text-align:right;"> 0 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:29:07 </td>
   <td style="text-align:left;"> InBedroomMaster </td>
   <td style="text-align:right;"> 1 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:29:16 </td>
   <td style="text-align:left;"> InBathroomMaster </td>
   <td style="text-align:right;"> 1 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:36:39 </td>
   <td style="text-align:left;"> InBathroomMaster </td>
   <td style="text-align:right;"> 1 </td>
  </tr>
</tbody>
</table>

---

exclude: true

<table>
 <thead>
  <tr>
   <th style="text-align:left;"> Timestamp </th>
   <th style="text-align:left;"> Sensor </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:01:44 </td>
   <td style="text-align:left;"> InLivingRoom </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:05:28 </td>
   <td style="text-align:left;"> InLivingRoom </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:26:25 </td>
   <td style="text-align:left;"> InKitchen </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:28:34 </td>
   <td style="text-align:left;"> <strong>OpenedGarageDoor</strong> </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:28:35 </td>
   <td style="text-align:left;"> InLivingRoom </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:28:58 </td>
   <td style="text-align:left;"> <strong>!OpenedGarageDoor</strong> </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:29:07 </td>
   <td style="text-align:left;"> InBedroomMaster </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:29:16 </td>
   <td style="text-align:left;"> InBathroomMaster </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:36:39 </td>
   <td style="text-align:left;"> InBathroomMaster </td>
  </tr>
</tbody>
</table>

---

<table>
 <thead>
  <tr>
   <th style="text-align:left;"> Timestamp </th>
   <th style="text-align:left;"> Sensor </th>
   <th style="text-align:left;"> Encoded </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:01:44 </td>
   <td style="text-align:left;"> InLivingRoom </td>
   <td style="text-align:left;"> A </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:05:28 </td>
   <td style="text-align:left;"> InLivingRoom </td>
   <td style="text-align:left;"> A </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:26:25 </td>
   <td style="text-align:left;"> InKitchen </td>
   <td style="text-align:left;"> E </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:28:34 </td>
   <td style="text-align:left;"> OpenedGarageDoor </td>
   <td style="text-align:left;"> G </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:28:35 </td>
   <td style="text-align:left;"> InLivingRoom </td>
   <td style="text-align:left;"> A </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:28:58 </td>
   <td style="text-align:left;"> ClosedGarageDoor </td>
   <td style="text-align:left;"> g </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:29:07 </td>
   <td style="text-align:left;"> InBedroomMaster </td>
   <td style="text-align:left;"> I </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:29:16 </td>
   <td style="text-align:left;"> InBathroomMaster </td>
   <td style="text-align:left;"> i </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2017-10-01 14:36:39 </td>
   <td style="text-align:left;"> InBathroomMaster </td>
   <td style="text-align:left;"> i </td>
  </tr>
</tbody>
</table>

---
layout: false

## Insert pauses for activity lulls

### Pauses

.center[
<table>
<thead>
  <tr>
   <th style="text-align:left;"> Type </th>
   <th style="text-align:left;"> Length </th>
   <th style="text-align:left;"> Character </th>
  </tr>
</thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Short </td>
   <td style="text-align:left;"> 15 to 60 minutes </td>
   <td style="text-align:left;"> <code>1</code> </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Medium </td>
   <td style="text-align:left;"> 1 to 3 hours </td>
   <td style="text-align:left;"> <code>2</code> </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Long </td>
   <td style="text-align:left;"> 3+ hours </td>
   <td style="text-align:left;"> <code>3</code> </td>
  </tr>
</tbody>
</table>
]

---

## Identify start of day

**Day Start Marker:** 5 sequential active events after 4am

<br>

.center[
<img src="gab-informs2017_files/figure-html/plot-activations-1.png" width="432" />
]

---
exclude: true
class: large

<br><br>

.center[
<table>
 <thead>
  <tr>
   <th style="text-align:left;"> Day </th>
   <th style="text-align:left;"> String </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> 2017-10-01 </td>
   <td style="text-align:left;"> <pre>AAEGAgIii</pre> </td>
  </tr>
</tbody>
</table>
]

---
layout: true
## Day Event Sequence
---

<br><br>
.ngram-example[IiIAEFfFfA2AE1EGAgEAIiIAEIiIiOiIGAgEIIIAE1AE2AEAAE1A1AAEIiIFfFfAEA2AEAE1AEEAAEAEAA1AIiI2AEIiI3IIiI2]

---

<br><br>
.ngram-example[.remark-code-line-highlighted[.highlight[IiIAE]]FfFfA2AE1EGAgEAIiIAEIiIiOiIGAgEIIIAE1AE2AEAAE1A1AAEIiIFfFfAEA2AEAE1AEEAAEAEAA1AIiI2AEIiI3IIiI2]

---
layout: false
### Activity Profiles: Inspired by text mining & document classification

Similar to document categorization [Cavnar1994]

- Counted and ranked *n*-grams in category documents

- Compared *n*-grams from unknown document to category profiles

**Activity Profile**

- For *m* day-event sequences:

- Count all *n*-grams from *n<sub>min</sub>* to *n<sub>max</sub>*
    - Calculate relative frequency *p* of each *n*-gram
    
---
### Gather _n_-grams from day event sequence

.center[
#### Event Sequence
`IiIAEF ... 3IIiI2`

#### Bigrams
`_I iI IA AE EF ... 3I II Ii iI 2_`

#### Trigrams
`_Ii iIA IAE AEF ... 3II IIi IiI I2_`<br><br> `...`
]

---
### Activity Profile: _bag of n-grams_ representation of day event sequences

.pull-left[
.center[**Day Activity Profile**]

<div style = 'font-size: 75%'>
<table>
 <thead>
  <tr>
   <th style="text-align:left;"> ngram </th>
   <th style="text-align:right;"> count </th>
   <th style="text-align:right;"> freq </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> AE </td>
   <td style="text-align:right;"> 15 </td>
   <td style="text-align:right;"> 0.02618 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> iI </td>
   <td style="text-align:right;"> 8 </td>
   <td style="text-align:right;"> 0.01396 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> EA </td>
   <td style="text-align:right;"> 7 </td>
   <td style="text-align:right;"> 0.01222 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Ii </td>
   <td style="text-align:right;"> 7 </td>
   <td style="text-align:right;"> 0.01222 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> IiI </td>
   <td style="text-align:right;"> 6 </td>
   <td style="text-align:right;"> 0.01047 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 1A </td>
   <td style="text-align:right;"> 5 </td>
   <td style="text-align:right;"> 0.00873 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> AEA </td>
   <td style="text-align:right;"> 5 </td>
   <td style="text-align:right;"> 0.00873 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2A </td>
   <td style="text-align:right;"> 4 </td>
   <td style="text-align:right;"> 0.00698 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 2AE </td>
   <td style="text-align:right;"> 4 </td>
   <td style="text-align:right;"> 0.00698 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> AA </td>
   <td style="text-align:right;"> 4 </td>
   <td style="text-align:right;"> 0.00698 </td>
  </tr>
</tbody>
</table>
</div>

.center[.small[450 *n*-grams]]
]

--

.pull-right[
.center[**Activity Profile**]

<div style = 'font-size: 75%'>
<table>
 <thead>
  <tr>
   <th style="text-align:left;"> ngram </th>
   <th style="text-align:right;"> count </th>
   <th style="text-align:right;"> freq </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> AE </td>
   <td style="text-align:right;"> 329 </td>
   <td style="text-align:right;"> 0.01541 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 1A </td>
   <td style="text-align:right;"> 283 </td>
   <td style="text-align:right;"> 0.01326 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Gg </td>
   <td style="text-align:right;"> 208 </td>
   <td style="text-align:right;"> 0.00974 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> EA </td>
   <td style="text-align:right;"> 193 </td>
   <td style="text-align:right;"> 0.00904 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> AA </td>
   <td style="text-align:right;"> 170 </td>
   <td style="text-align:right;"> 0.00796 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> iI </td>
   <td style="text-align:right;"> 168 </td>
   <td style="text-align:right;"> 0.00787 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> 1AE </td>
   <td style="text-align:right;"> 146 </td>
   <td style="text-align:right;"> 0.00684 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Ff </td>
   <td style="text-align:right;"> 139 </td>
   <td style="text-align:right;"> 0.00651 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> A1 </td>
   <td style="text-align:right;"> 134 </td>
   <td style="text-align:right;"> 0.00628 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> A1A </td>
   <td style="text-align:right;"> 132 </td>
   <td style="text-align:right;"> 0.00618 </td>
  </tr>
</tbody>
</table>
</div>

.center[.small[7,840 *n*-grams]]
]

---
layout: false
### Use KL Distance to measure distance between Activity Profiles

Comparing Activity Profile *P* to Daily Activity Profile *Q*

Kullback-Leibler Distance (Symmetric)

`$$D(P \Vert Q) = \sum_{x \in \chi} \left( \left(P(x) - Q(x)\right) \log \frac{P(x)}{Q(x)}\right)$$`

with *back-off*

`$$\begin{align}
P(t_k, d_j) &= \begin{cases} \beta P(t_k \vert d_j) & \text{if } t_k \text{ occurs in document } d_j \\ \epsilon & \text{otherwise} \end{cases} \\
\beta &= 1 - \sum_{k \not\in d_i, \;k \in P} \epsilon \\
\end{align}$$`

.footnote[
Bigi 2003
]

???

KL divergence is a measure of the difference between two probability distributions over the same event space

KLJ or KL distance is symmetric version

---
layout: false
## Identifying Occupants Using Activity Profiles

1. Do Activity Profiles capture occupant's unique patterns?

2. How many days of activity are needed?

3. Can we use a activity profiles as baseline as occupant ages?

---

background-image: url(images/profile-comparisons/occupant-profiles.svg)
background-position: 50% 125%
background-size: 90%
### Build Activity Profiles for each occupant from _N_ day event sequences

---

background-image: url(images/profile-comparisons/comparison-profiles2.svg)
background-position: 50% 125%
background-size: 90%
### Compare unknown single-day profile with all Activity Profiles to predict occupant

---
layout: true
## Experimental Setup
---

- 196 total days of data

- _How well do Activity Profiles work?_

- Build profile for each occupant from _N_ training days
    
    - Use profiles to identify occupant of test day
    
    - 7-fold modified cross validation:
        - 28 testing days (~ 1 month) per fold
        - 168 training days available for profile
    
- _How many days are needed for quality Activity Profiles?_

- *N* profile days:
    
        - From 14 days to 3 months

---
layout: false
### Testing days from contiguous days

<img src="gab-informs2017_files/figure-html/folds-1-1.png" width="864" />

---

### Profiles built from _N_ sampled days

<img src="gab-informs2017_files/figure-html/folds-2-1.png" width="864" />
    
---
### Samples are repeated for equal number of profile day trials

Profiles: 168 days (~6 months)

Test: 28 days

<table>
 <thead>
  <tr>
   <th style="text-align:right;"> Profile Days </th>
   <th style="text-align:right;"> Months </th>
   <th style="text-align:right;"> Iterations </th>
   <th style="text-align:right;"> Reps </th>
   <th style="text-align:right;"> Total Trials </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:right;"> 14 </td>
   <td style="text-align:right;"> 0.5 </td>
   <td style="text-align:right;"> 12 </td>
   <td style="text-align:right;"> 1 </td>
   <td style="text-align:right;"> 12 </td>
  </tr>
  <tr>
   <td style="text-align:right;"> 28 </td>
   <td style="text-align:right;"> 1.0 </td>
   <td style="text-align:right;"> 6 </td>
   <td style="text-align:right;"> 2 </td>
   <td style="text-align:right;"> 12 </td>
  </tr>
  <tr>
   <td style="text-align:right;"> 42 </td>
   <td style="text-align:right;"> 1.5 </td>
   <td style="text-align:right;"> 4 </td>
   <td style="text-align:right;"> 3 </td>
   <td style="text-align:right;"> 12 </td>
  </tr>
  <tr>
   <td style="text-align:right;"> 56 </td>
   <td style="text-align:right;"> 2.0 </td>
   <td style="text-align:right;"> 3 </td>
   <td style="text-align:right;"> 4 </td>
   <td style="text-align:right;"> 12 </td>
  </tr>
  <tr>
   <td style="text-align:right;"> 70 </td>
   <td style="text-align:right;"> 2.5 </td>
   <td style="text-align:right;"> 2 </td>
   <td style="text-align:right;"> 6 </td>
   <td style="text-align:right;"> 12 </td>
  </tr>
  <tr>
   <td style="text-align:right;"> 84 </td>
   <td style="text-align:right;"> 3.0 </td>
   <td style="text-align:right;"> 2 </td>
   <td style="text-align:right;"> 6 </td>
   <td style="text-align:right;"> 12 </td>
  </tr>
</tbody>
</table>

---
layout: false
## Results

<table>
 <thead>
  <tr>
   <th style="text-align:left;">  </th>
   <th style="text-align:right;"> 14 Days </th>
   <th style="text-align:right;"> 28 Days </th>
   <th style="text-align:right;"> 42 Days </th>
   <th style="text-align:right;"> 56 Days </th>
   <th style="text-align:right;"> 70 Days </th>
   <th style="text-align:right;"> 84 Days </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Accuracy </td>
   <td style="text-align:right;"> 0.944 </td>
   <td style="text-align:right;"> 0.959 </td>
   <td style="text-align:right;"> 0.962 </td>
   <td style="text-align:right;"> 0.960 </td>
   <td style="text-align:right;"> 0.959 </td>
   <td style="text-align:right;"> 0.954 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Precision </td>
   <td style="text-align:right;"> 0.949 </td>
   <td style="text-align:right;"> 0.962 </td>
   <td style="text-align:right;"> 0.965 </td>
   <td style="text-align:right;"> 0.962 </td>
   <td style="text-align:right;"> 0.962 </td>
   <td style="text-align:right;"> 0.957 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Recall </td>
   <td style="text-align:right;"> 0.944 </td>
   <td style="text-align:right;"> 0.959 </td>
   <td style="text-align:right;"> 0.962 </td>
   <td style="text-align:right;"> 0.960 </td>
   <td style="text-align:right;"> 0.959 </td>
   <td style="text-align:right;"> 0.954 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> F Meas. </td>
   <td style="text-align:right;"> 0.944 </td>
   <td style="text-align:right;"> 0.959 </td>
   <td style="text-align:right;"> 0.962 </td>
   <td style="text-align:right;"> 0.960 </td>
   <td style="text-align:right;"> 0.959 </td>
   <td style="text-align:right;"> 0.954 </td>
  </tr>
</tbody>
</table>

---
## Average Accuracy

<img src="gab-informs2017_files/figure-html/plot-summary-1.png" width="864" />

---
## Average Accuracy by Occupant

<img src="gab-informs2017_files/figure-html/plot-house-summary-1.png" width="864" />

---
layout: false
### First 42 days as baseline profile<br>compared with remaining 154 days

<img src="gab-informs2017_files/figure-html/plot-self-sim-1.png" width="1008" />

---
## Summary

**Activity Profiles**

- Accurately characterize smart home occupant daily activities

- Useful for long-term monitoring of older adults as they age

- Use simple, passive sensors emebedded in home environment

- Do not require expensive activity labelling for training

Future Work

Questions? <gadenbuie@mail.usf.edu>

???

Future work: clustering for unsupervised activity discovery

Limitations: obviously, no labels, etc.

---
class: small
## References

**Wouters** (2017). Handbook of Smart Homes, Health Care and Well-Being

**Christensen et al.** (2009). doi:[10.1016/S0140-6736(09)61460-4](http://linkinghub.elsevier.com/retrieve/pii/S0140673609614604)

**Federal Interagency Forum on Aging-Related Statistics** (2012). Older Americans 2012

**Bodenheimer et al.** (2009). doi:[10.1377/hlthaff.28.1.64](https://doi.org/10.1377/hlthaff.28.1.64)

**Scommegna** (2016). Today’s Research on Aging

**U. S. Census Bureau** (2017). <https://factfinder.census.gov/bkmk/table/1.0/en/ACS/15_5YR/S0103/0100000US>

**Centers for Disease Control and Prevention** (2017). <https://www.cdc.gov/healthyplaces/terminology.htm>

**Mather et al.** (2015). Aging in the United States

**Cavnar & Trenkle** (1994). N Proceedings of SDAIR-94, 3rd Annual Symposium on Document Analysis and Information Retrieval

**Bigi** (2003). Advances in Information Retrieval