On Foundations of Estimation for Nonparametric Regression With Continuous Optimization

The aim of parametric regression models like linear regression and nonlinear regression are to produce a reasonable relationship between response and independent variables based on the assumption of linearity and predetermined nonlinearity in the regression parameters by finite set of parameters. Nonparametric regression techniques are widely-used statistical techniques, and they not only relax the assumption of linearity in the regression parameters, but they also do not need a predetermined functional form as nonlinearity for the relationship between response and independent variables. It is capable of handling higher dimensional problem and sizes of sample than regression that considers parametric models because the data should provide both the model building and the model estimates. For this purpose, firstly, PRSS problems for MARS, ADMs, and CR will be constructed. Secondly, the solution of the generated problems will be obtained with CQP, one of the famous methods of convex optimization, and these solutions will be called CMARS, CADMs, and CKR, respectively.

Big Data Analytics and Models

When uncertainty is regarded as a surprise and an event in the minds, it can be said that individuals can change the future view. Market, financial, operational, social, environmental, institutional and humanitarian risks and uncertainties are the inherent realities of the modern world. Life is suffused with randomness and volatility; everything momentous that occurs in the illustrious sweep of history, or in our individual lives, is an outcome of uncertainty. An important implication of such uncertainty is the financial instability engendered to the victims of different sorts of perils. This chapter is intended to explore big data analytics as a comprehensive technique for processing large amounts of data to uncover insights. Several techniques before big data analytics like financial econometrics and optimization models have been used. Therefore, initially these techniques are mentioned. Then, how big data analytics has altered the methods of analysis is mentioned. Lastly, cases promoting big data analytics are mentioned.

Big Data and Clustering Techniques

Big data has become a chief strength of innovation across academics, governments, and corporates. Big data comprises massive sensor data, raw and semi-structured log data of IT industries, and the exploded quantity of data from social media. Big data needs big storage, and this volume makes operations such as analytical operations, process operations, retrieval operations very difficult and time consuming. One way to overcome these difficult problems is to have big data clustered in a compact format. Thus, this chapter discusses the background of big data and clustering. It also discusses the various application of big data in detail. The various related work, research challenges of big data, and the future direction are addressed in this chapter.

Urban Spatial Data Computing

Urban spatial data is the source of information in analysing risks due to natural disaster, evacuation planning, risk mapping and assessments, etc. Global positioning system (GPS) is a satellite-based technology that is used to navigate on earth. Geographical information system (GIS) is a software system that facilitates software services to mankind in various application domains such as agriculture, ecology, forestry, geomorphology analysis in earthquake and landslides, laying of underground water pipe connection and demographic studies like population migration, urban settlements, etc. Thus, spatial and temporal relations of real-time activities can be analysed to predict the future activities like predicting places of interest. Time analysis of such activities helps in personalisation of activities or development of recommendation systems, which could suggest places of interest. Thus, GPS mapping with data analytics using GIS would pave way for commercial and business development in large scale.

Examining Visitors' Characteristics and Behaviors in Tourist Destinations Through Mobile Phone Users' Location Data

This chapter uses Mobile Kukan Toukei™ (mobile spatial statistics) to collect the location data of mobile phone users in order to count the number of visitors at specific tourist destinations and examine their characteristics. Mobile Kukan Toukei is statistical population data created by an operational data of mobile phone networks. It is possible to estimate the population structure of a region by gender, age, and residence using this service of the company. The locations and characteristics of the individuals obtained herein are derived through a non-identification process, aggregation processing, and concealment processing. Therefore, it is impossible to identify specific individuals. This chapter attempts to identify the number of visitors in different periods and their characteristics based on the location data of mobile phone users collected by the mobile phone company. In addition, it also attempts to demonstrate an alternative method to more accurately infer the number of visitors in specific areas.

Analytics and Technology for Practical Forecasting

With the importance of forecasting in businesses, a wide variety of methods and tools has been developed over the years to automate the process of forecasting. However, an unintended consequence of this tremendous advancement is that forecasting has become more and more like a black box function. Thus, a primary goal of this chapter is to provide a clear understanding of forecasting in any application contexts with a systematic procedure for practical forecasting through step-by-step examples. Several methods are presented, and the authors compare results to what were the typical forecasting methods including regression and time series in different software technologies. Three case studies are presented: simple supply forecasting, homicide forecasting, and demand forecasting for sales from Walmart.

Technologies for Handling Big Data

In today's world, every time we connect phone to internet, pass through a CCTV camera, order pizza online, or even pay with credit card to buy some clothes, we generate data and that “ocean of data” is popularly known as big data. The amount of data that's being created and stored on a universal level is almost inconceivable, and it just keeps growing. The amount of data we create is doubled every year. Big data is a critical concept that integrates all kinds of data and plays an important role for strategic intelligence for any modern company. The importance of big data doesn't revolve around how much data you have, but what you do with it. Big data is now the key for competition and growth for new startups, medium, and big enterprises. Scientific research is now on boom using big data. For the astronomers, Sloan Digital Sky Survey has become a central resource. Big data has the potential to revolutionize research and education as well. The aim of this chapter is to discuss the technologies that are pertinent and essential for big data.

Call Masking

The phenomenon of call masking and other related infractions have assumed frightening dimension in Nigeria. Apart from depriving the government and telecoms companies of huge revenue, the sharp practices also constitute a security threat to the nation. In a bid to curb the menace, the Nigerian Communications Commission, the industry regulator, had to suspend six interconnect exchange licenses in February 2018 and bar 750,000 lines belonging to 13 operators from the national network suspected to have been involved in the criminal act. However, in spite of the measures taken by NCC, the sharp practices have continued unabated. It is against this backdrop that this chapter proffers solutions and recommends ways to nip the infractions in the bud and save the telecoms industry from imminent collapse.

Data-Driven Stochastic Optimization for Transportation Road Network Design Under Uncertainty

A data-driven stochastic program for bi-level network design with hazardous material (hazmat) transportation is proposed in this chapter. In order to regulate the risk associated with hazmat transportation and minimize total travel cost on interested area under stochasticity, a multi-objective stochastic optimization model is presented to determine generalized travel cost for hazmat carriers. Since the bi-level program is generally non-convex, a data-driven bundle method is presented to stabilize solutions of the proposed model and reduce relative gaps between iterations. Numerical comparisons are made with existing risk-averse models. The results indicate that the proposed data-driven stochastic model becomes more resilient than others in minimizing total travel cost and mitigating risk exposure. Moreover, the trade-offs among maximum risk exposure, generalized travel costs, and maximum equitable risk spreading over links are empirically investigated in this chapter.

An Overview of Methodologies and Challenges in Sentiment Analysis on Social Networks

The expeditious increase in the adoption of social media over the last decade, determining and analyzing the attitude and opinion of masses related to a particular entity, has gained quite an importance. With the landing of the Web 2.0, many internet products like blogs, community chatrooms, forums, microblog are serving as a platform for people to express themselves. Such opinion is found in the form of messages, user-comments, news articles, personal blogs, tweets, surveys, status updates, etc. With sentiment analysis, it is possible to eliminate the need to manually going through each and every user comment by focusing on the contextual polarity of the text. Analyzing the sentiments could serve a number of applications like advertisements, recommendations, quality analysis, monetization provided on the web services, real-time analysis of data, analyzing notions related to candidates during election campaign, etc.