Enhancing Activity Recognition in Smart Homes Using Feature Induction

Nair, Naveen; Ramakrishnan, Ganesh; Krishnaswamy, Shonali

doi:10.1007/978-3-642-23544-3_31

Cited by 6 publications

(5 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the UA data, we compared our results with eight other approaches: (a) standard HMM [22], (b) Branch and Bound structure learning assisted HMM model construction (B&B HMM), where the rules learned by Aleph [24] (an ILP system which learns definite rules from examples) for each activity determine the HMM emission structure, (c) greedy feature induction assisted HMM approach (Greedy FIHMM) [20], (d) StructSVM approach [28], (e) Conditional Random Field (CRF) [10], (f) Conditional Random Field with Feature Induction (FICRF) [16,15], (g) RELHKL (without considering transitions) [9] Greedy FIHMM and FICRF use conjunctions of basic features as emission features. In contrast to greedy feature induction approaches, RELHKL, and StructHKL find the feature conjunctions efficiently and optimally.…”

Section: Methodsmentioning

confidence: 99%

“…First, we discuss greedy feature induction approaches for sequential data and then discuss an optimal feature induction approach that works for binary classification problems. Mc-Callum [16] as well as our prior work [20], propose feature induction methods that iteratively construct feature conjunctions that increase an objective. These approaches start with an initial set of features, and at each step, consider a set of candidate features (conjunctions or atomic).…”

Section: Learning Relationships As Featuresmentioning

confidence: 99%

“…The user activity at a particular time forms the hidden state or label. The history of sensor readings and (manually) annotated activities can be used to train prediction models which could later be used to predict activities based on sensor observations [29,20]. Hidden Markov Models (HMM) [22], Conditional Random Fields (CRF) [10] and Support Vector Machines on Structured Output Spaces (StructSVM) [27,28] are the most popular approaches to sequence labeling.…”

Section: Introductionmentioning

confidence: 99%

“…Recent works have shown that learning the relational structure between input features improves the efficiency of sequence labeling models [14,20]. For example, in activity recognition settings, one could expect certain combinations of (sensor) readings to be directly indicative of certain activities (for example, microwave and platesCupboard sensors firing within a short time period could be a strong indicator of the activity preparingDinner).…”

Section: Introductionmentioning

confidence: 99%

“…An outline of our optimal learning approaches for the useful feature categories is as follows: to begin with, we learn features that are conjunctions of basic inputs at a single sequence position, which is the feature space explored in [20] and [16]. We propose a Hierarchical Kernels-based approach for discovering features from this space.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Learning Discriminative Relational Features for Sequence Labeling

Nair¹,

Nagesh²,

Ramakrishnan³

2017

Preprint

Self Cite

View full text Add to dashboard Cite

Discovering relational structure between input features in sequence labeling models has shown to improve their accuracy in several problem settings. However, the search space of relational features is exponential in the number of basic input features. Consequently, approaches that learn relational features, tend to follow a greedy search strategy. In this paper, we study the possibility of optimally learning and applying discriminative relational features for sequence labeling. For learning features derived from inputs at a particular sequence position, we propose a Hierarchical Kernels-based approach (referred to as Hierarchical Kernel Learning for Structured Output Spaces -StructHKL). This approach optimally and efficiently explores the hierarchical structure of the feature space for problems with structured output spaces such as sequence labeling. Since the StructHKL approach has limitations in learning complex relational features derived from inputs at relative positions, we propose two solutions to learn relational features namely, (i) enumerating simple component features of complex relational features and discovering their compositions using StructHKL and (ii) leveraging relational kernels, that compute the similarity between instances implicitly, in the sequence labeling problem. We perform extensive empirical evaluation on publicly available datasets and record our observations on settings in which certain approaches are effective.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Learning Relationships As Featuresmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Learning Discriminative Relational Features for Sequence Labeling

Nair¹,

Nagesh²,

Ramakrishnan³

2017

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

Probing the Space of Optimal Markov Logic Networks for Sequence Labeling

Nair

Nagesh

Ramakrishnan

2013

Inductive Logic Programming

View full text Add to dashboard Cite

Rule Ensemble Learning Using Hierarchical Kernels in Structured Output Spaces

Nair

Saha

Ramakrishnan

et al. 2021

AAAI

View full text Add to dashboard Cite

The goal in Rule Ensemble Learning (REL) is simultaneous discovery of a small set of simple rules and their optimal weights that lead to good generalization. Rules are assumed to be conjunctions of basic propositions concerning the values taken by the input features. It has been shown that rule ensembles for classification can be learnt optimally and efficiently using hierarchical kernel learning approaches that explore the exponentially large space of conjunctions by exploiting its hierarchical structure. The regularizer employed penalizes large features and thereby selects a small set of short features. In this paper, we generalize the rule ensemble learning using hierarchical kernels (RELHKL) framework to multi class structured output spaces. We build on the StructSVM model for sequence prediction problems and employ a ρ-norm hierarchical regularizer for observation features and a conventional 2-norm regularizer for state transition features. The exponentially large feature space is searched using an active set algorithm and the exponentially large set of constraints are handled using a cutting plane algorithm. The approach can be easily extended to other structured output problems. We perform experiments on activity recognition datasets which are prone to noise, sparseness and skewness. We demonstrate that our approach outperforms other approaches.

show abstract

Enhancing Activity Recognition in Smart Homes Using Feature Induction

Cited by 6 publications

References 4 publications

Learning Discriminative Relational Features for Sequence Labeling

Learning Discriminative Relational Features for Sequence Labeling

Probing the Space of Optimal Markov Logic Networks for Sequence Labeling

Rule Ensemble Learning Using Hierarchical Kernels in Structured Output Spaces

Contact Info

Product

Resources

About