Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Nanoscale chemical surface characterization of four different types of dental pulp-capping materials

Currently, for content-based recommendations, semantic analysis of text from webpages seems to be a major problem. In this research, we present a semantic web content mining approach
for recommender systems in online shopping. The methodology is based on two major phases. The first phase is the semantic preprocessing of textual data using the combination of a
developed ontology and an existing ontology. The second phase uses the Naive Bayes algorithm to make the recommendations. The output of the system is evaluated using precision, recall
and f-measure. The results from the system showed that the semantic preprocessing improved the recommendation accuracy of the recommender system by 5.2% over the existing
approach. Also, the developed system is able to provide a platform for content-based recommendation in online shopping. This system has an edge over the existing recommender
approaches because it is able to analyze the textual contents of users feedback on a product in order to provide the necessary product recommendation.

Semantic Web mining for Content-Based Online Shopping Recommender Systems

Visually Impaired People (VIP) have to move in the highly dense society alone. In real world situation, they have to overcome more obstacles, hurdles and traffic while they navigate indoor and outdoor. Even more sophisticated technology cannot help those people for their convenience navigation and utility. The virtual vision architecture is composed of different subsystem. This architecture includes Head Obstacle Detection system, Tail Obstacle Detection system (TOD), Positioning and Location System, Alerting and Notification System, Information Management System (IMS) and Speech Recognizer Engine. IMS consists of Selenium web driver that is used to retrieve the latest information from various web servers. It is a newly proposed method to communicate with the existing web server. A TOD system is capable of monitoring the moving objects that comes behind the VIP. The proposed idea includes three methods for calculating the distance of the moving object. The speed is calculated from the distance. Based on the speed, the walking direction of the VIP is adjusted to avoid an accident.

Virtual Vision Architecture for VIP in Ubiquitous Computing

The investigation of surfaces and thin films is of particular interest in current research as it provides a basis for a multiplicity of applications such as waveguides, sensors, solar cells and optoelectronics. The origins of complex phenomena on surfaces and in thin films can be revealed by applying angle-resolved spectroscopy in two dimensions: the angle of incidence is scanned while analyzing the full emission spectrum in terms of exit angle and emission energy. The interface of a metal layer and a dielectric can support collective electron plasma resonances, i.e. surface plasmon polaritons, which are accompanied by giant field enhancement while propagating along the interface. We characterize the Kretschmann and the Otto coupling configuration in terms of their coupling efficiency and their impact on the surface plasmon resonance as a function of wave-vector. Although being commonly considered as equivalent in terms of plasmonic coupling, we identify differing dependencies of their respective coupling efficiency on the coupling layer thickness and the excitation wavelength which is fundamental for sensing applications. Provided that a metal layer is embedded in a symmetric cladding in terms of its dielectric function and the film thickness is reduced to the order of λ/10, modes from both interfaces can couple and propagate as long-range surface waves. Surprisingly, even intrinsically absorbing films support low-loss surface waves, whose propagation length can become arbitrarily long in the limit of vanishing film thickness. This phenomenon requires only that the material’s dielectric function be predominantly imaginary over that particular range of optical frequencies. Furthermore we show that the orientation of transition dipole moments inside thin monolayer films of effective media that contain oriented CdSe nano-platelets can be determined by applying k-space spectroscopy. Thus we determine electronic and dielectric contributions to the emission anisotropy and reveal the intrinsic nature of the directionality in the emission. We show that this phenomenon is related to the anisotropy of the electronic Bloch states that govern the transition dipole moment of the exciton. Beyond the linear investigation of surfaces and thin films, 2D-k-space spectroscopy can provide an insight into the principles of nonlinear wave-mixing interactions. The role of surface plasmons in second harmonic generation, whether they act as field-enhancing catalysts or as quasiparticles converted in the interaction can be revealed by k-space spectroscopy: by way of the signature in k-space, we identify a nonlinear interaction where two surface plasmons annihilate to create a second-harmonic photon as well as the interaction of a plasmon and a photon by virtue of a degenerate three-wave mixing process. We analyze the intrinsic origin of surface plasmon enhanced second harmonic generation in metal films by comparing the absolute nonlinear yield in attenuated internal reflection configurations to theoretical calculations based on the hydrodynamic model. A first estimation of the nonlinear parameters in the hydrodynamic model is given and the contributions of the bulk and surface source are determined, showing that the in-plane surface source is stronger than predicted. For absorbing thin films however, we report the first evidence of field enhancement and long-range surface wave enhanced second harmonic generation. Here, we identify the out-of-plane surface source to have the strongest contribution to the second harmonic yield.As the nonlinear susceptibility of a material can greatly increase if the probing frequency approaches an absorption resonance, absorbing materials can indeed be considered as low-loss optical media for doing surface-wave optics in the nonlinear regime. We show further, that, in contrast to the isotropic linear absorption, the two-photon absorption in oriented nano-platelets is highly anisotropic. This transition dipole orientation is dependent on the probabilities of the involved processes and their selection rules. We demonstrate that an additional silver layer covered with SiO2 enables surface plasmon enhanced excitation of oriented nano-platelets, boosting the photoluminescent emission which is highly directed through coupling to the plasmonic mode. The combination of TPA and the plasmonic resonance even leads to further concentration of the absorption range as a function of excitation wave-vector. In summary, this work has shown that 2D-k-space spectroscopy –as applied to solid surfaces, thin films and nano-particles– provides insight into the intrinsic material properties, as well as the surfacewave and radiation phenomena supported by these structures.

Investigations of nonlinear surface-wave phenomena and anisotropic photoluminescence of nano-film structures via two-dimensional k-space spectroscopy

In recent times, big data analytics has become a major trend in catering data queries that has been growing dramatically. The present paper gives a brief description of latest happenings of Big Data analytics. A case study using Spark is given as an example. The paper gives the importance of cloud computing in Big data paradigm.

/pdf/foundations-to-frontiers-of-big-data-analytics-n5b5opo29r.pdf

Foundations to frontiers of big data analytics

Mining of Bilingual Indian Web Documents

Two major technological developments — cellular and internet communications — that have significantly changed the outlook of society throughout the world and more so in the Indian context. Usage and interaction through cell-phones and internet are so popular in different parts of India without language or dialect barrier. Education is one discipline where Indian input has been phenomenal both in terms of students and institutions and soon the switch over to online education and learning will take place to cater to different levels and 24/7 availability. The present study looks at issues in having a common web-based content for developing the courses. Towards this object and subject issues are taken and based on a generic pixel based approach, a common reference content prediction is looked at. Case studies on ‘circle’ — object -,‘internet’ — a subject and ‘magnetism’ — a part of science and physics demonstrate applicability and versatility in handling different document types.

Complexities in developing multilingual on-line courses in the Indian context

Internet and mobile computing have become a major societal force in that down-to-earth issues are being addressed and sorted out whether they relate to online shopping or securing driving information in unknown places. Here the major concern of communication is that the Web content should reach the user in a short period of time. So information extraction needs to be at a basic level and easier to implement without depending on any major software. The present study focuses on extraction of information from the available text and media-type data after it is converted into digital form. The approach uses the basic pixel map representation of data and converting them through numerical means, so that issues of language, text script and format do not pose problems. With the numerically converted data, key clusters similar to keywords used in any search method are developed and content is extracted through different approaches making it computation-intensive for easiness. One approach is that statistical features of the images are extracted from the pixel map of the image. The extracted features are presented to the fuzzy clustering algorithm. The similarity metric being Euclidean distance and the accuracy is compared and presented. The concept of ambiguity is introduced in the paper, by comparing objects like ‘computer,’ which have explicit content representation possible to an abstract subject like ‘soft-computing,’ where vagueness and ambiguity are possible in representation. With this as the objective, the approach used for content extraction is compared and how within certain bounds it could be possible to extract the content.

Mining Ambiguities Using Pixel-Based Content Extraction

Web pages nowadays have different forms and types of content. When the Web content is considered, they are in the form of pictures, videos, audio files, and text files in different languages. The content can be multilingual, heterogeneous, and unstructured. The mining should be independent of the language and software. Statistical features of the images are extracted from the pixel map of the image. The extracted features are presented to the fuzzy clustering algorithm (FCM) and Gath–Geva algorithm. The similarity metric being Euclidean distance and Gaussian distance, respectively. The accuracy is compared and presented.

Arun Rajaraman

Papers

Foundations to frontiers of big data analytics

Mining of Bilingual Indian Web Documents

Complexities in developing multilingual on-line courses in the Indian context

Mining Ambiguities Using Pixel-Based Content Extraction

Clustering for Knowledgeable Web Mining