Data Mining - Concepts and Techniques.

Data Mining Practical Machine Learning Tools and Techniques

As we are moving towards the Internet of Things (IoT), the number of sensors deployed around the world is growing at a rapid pace. Market research has shown a significant growth of sensor deployments over the past decade and has predicted a significant increment of the growth rate in the future. These sensors continuously generate enormous amounts of data. However, in order to add value to raw sensor data we need to understand it. Collection, modelling, reasoning, and distribution of context in relation to sensor data plays critical role in this challenge. Context-aware computing has proven to be successful in understanding sensor data. In this paper, we survey context awareness from an IoT perspective. We present the necessary background by introducing the IoT paradigm and context-aware fundamentals at the beginning. Then we provide an in-depth analysis of context life cycle. We evaluate a subset of projects (50) which represent the majority of research and commercial solutions proposed in the field of context-aware computing conducted over the last decade (2001-2011) based on our own taxonomy. Finally, based on our evaluation, we highlight the lessons to be learnt from the past and some possible directions for future research. The survey addresses a broad range of techniques, methods, models, functionalities, systems, applications, and middleware solutions related to context awareness and IoT. Our goal is not only to analyse, compare and consolidate past research work but also to appreciate their findings and discuss their applicability towards the IoT.

/pdf/context-aware-computing-for-the-internet-of-things-a-survey-3nibvvc921.pdf

Context Aware Computing for The Internet of Things: A Survey

Matrix Factorization Techniques for Recommender Systems

We quantitatively investigate how machine learning models leak information about the individual data records on which they were trained. We focus on the basic membership inference attack: given a data record and black-box access to a model, determine if the record was in the model's training dataset. To perform membership inference against a target model, we make adversarial use of machine learning and train our own inference model to recognize differences in the target model's predictions on the inputs that it trained on versus the inputs that it did not train on. We empirically evaluate our inference techniques on classification models trained by commercial "machine learning as a service" providers such as Google and Amazon. Using realistic datasets and classification tasks, including a hospital discharge dataset whose membership is sensitive from the privacy perspective, we show that these models can be vulnerable to membership inference attacks. We then investigate the factors that influence this leakage and evaluate mitigation strategies.

/pdf/membership-inference-attacks-against-machine-learning-models-2tzchejf91.pdf

Membership Inference Attacks Against Machine Learning Models

GPS-equipped taxis can be considered as pervasive sensors and the large-scale digital traces produced allow us to reveal many hidden facts about the city dynamics and human behaviors. In this paper we present a novel GPS-based taxi system which can detect ongoing anomalous passenger delivery behaviors leveraging our proposed iBOAT method. To achieve real time monitoring, we reduce the response time of iBOAT by more than five times with an inverted index mechanism adopted. We evaluate the effectiveness of the system with large scale real life taxi GPS records while serving 200,000 taxis. With this system, we obtain about 0.44 million anomalous trajectories out of 7.35 million taxi delivery trips, which correspond to 7600 taxis' GPS records in one month time in the city of Hangzhou, China. Through further analysis of these anomalous trajectories, we observe that: (1) Over 60 % of the anomalous trajectories are "detours" that travel longer distances and time than normal trajectories; (2) The average trip length of drivers with high-detour tendency is 20 % longer than that of normal drivers; (3) The length of anomalous sub-trajectories is usually less than a third of the entire trip, and they tend to begin in the first two thirds of the journey; (4) Although longer distance results in a greater taxi fare, a higher tendency to take anomalous detours does not result in higher monthly revenue; and (5) Taxis with a higher income usually spend less time finding new passengers and deliver them in faster speed.

Real Time Anomalous Trajectory Detection and Analysis

Device-free human intrusion detection holds great potential and multiple challenges for applications ranging from asset protection to elder care. In this paper, leveraging the fine-grained Channel State Information (CSI) in commodity WiFi devices, we design and implement an adaptive and robust human intrusion detection system, called AR-Alarm. By utilizing a robust feature and self-adaptive learning mechanism, AR-Alarm achieves real-time intrusion detection in different environments without calibration efforts. To further increase the system robustness, we propose a few novel methods to distinguish real human intrusion from object motion in daily life such as object dropping, curtain swinging and pets moving. As demonstrated in the experiments, AR-Alarm achieves a high detection rate and low false alarm rate.

AR-Alarm: An Adaptive and Robust Intrusion Detection System Leveraging CSI from Commodity Wi-Fi

A new context management framework, which supports the interaction between different domain context managers, is proposed. Two different context producer-consumer patterns are defined to support context-aware applications. To deal with the mobile entity problem in cross-domain context sharing, a transparent query mechanism that enables applications to obtain context information about mobile entities from remote domains is presented. The capability of our framework is demonstrated through a user scenario.

The architecture design of a cross-domain context management system

Recent years have witnessed a trend of monitoring human respiration using Channel State Information (CSI) retrieved from commodity WiFi devices. Existing approaches essentially leverage signal propagation in a Line-of-Sight (LoS) setting to achieve good performance. However, in real-life environments, LoS can be easily blocked by furniture, home appliances and walls. This paper presents a novel smartphone-based system named WiPhone, aiming to robustly monitor human respiration in NLoS settings. Since a smartphone is usually carried around by one subject, leveraging directly-reflected CSI signals in LoS becomes infeasible. WiPhone exploits ambient reflected CSI signals in a Non-Line-of-Sight (NLoS) setting to quantify the relationship between CSI signals reflected from the environment and a subject's chest displacement. In this way, WiPhone successfully turns ambient reflected signals which have been previously considered "destructive" into beneficial sensing capability. CSI signals obtained from smartphone are usually very noisy and may scatter over different sub-carriers. We propose a density-based preprocessing method to extract useful CSI amplitude patterns for effective respiration monitoring. We conduct extensive experiments with 8 subjects in a real home environment. WiPhone achieves a respiration rate error of 0.31 bpm (breaths per minute) on average in a range of NLoS settings.

https://hal.archives-ouvertes.fr/hal-03363396/document

WiPhone: Smartphone-based Respiration Monitoring Using Ambient Reflected WiFi Signals

Offline event marketing invites people to participate in a sponsored gathering, thus allowing marketers to have face-to-face, direct, and close contact with their current and potential customers. This paper presents a framework that supports marketers in improving marketing effectiveness by carefully selecting invitees to such sponsored offline events by leveraging location-based social networks. In particular, we first transform the participant selection task into a combinatorial optimization problem. Second, we propose a marketing effect quantitative model that considers the distance and overlapping social influence. Third, we introduce algorithms to determine a participant team that can maximize the marketing effect while fulfilling the scale and item coverage constraints. We finally evaluate the effectiveness of the framework and validate the proposed marketing effect of the quantitative model with real-world data.

Daqing Zhang

Papers

Real Time Anomalous Trajectory Detection and Analysis

AR-Alarm: An Adaptive and Robust Intrusion Detection System Leveraging CSI from Commodity Wi-Fi

The architecture design of a cross-domain context management system

WiPhone: Smartphone-based Respiration Monitoring Using Ambient Reflected WiFi Signals

Participant Selection for Offline Event Marketing Leveraging Location-Based Social Networks