This research paper is about to basic conceptswith adaptive smart platform for intelligent software relating to big data [1]. It a well-defining attempt to integrate that make up big data, what metrics clarify the size and each characteristics of big data, and what methods and technologies consist in harness of using intelligent software for adaptive smart platform. Big data are the main concept to appeared in new technological developments [2]. In addition, big data-driven intelligent software is read to acceptance of their concept rapidly, and also, adaptive smart platform advance in high technology have increasingly complex and large-scale computer and communication systems[3]. Such as, Internet of things has exponentially increased the scale and complexity of existing their systems that adaptation is thus an imperative property for Internet of Everything. After all, main concept of this research paper is to bring 4th industrial revolution by big data environment, that is charge of commercial efforts by large intelligent software and smart platform. So to speak, big data focuses on ‘Prediction’, ‘Adaptation’, ‘Well-defined Structure’, ‘Analysis and Visualization’.

This paper is organized as follows. In Section 2 details the Adaptive Smart Platform. Then Intelligent Software in Section 3. Also, Section 4 presents Big Data-based Adaptive Smart Platform. Finally, Section 5 concludes the paper by briefly summarizing the main points.

The state of infrastructures and their technology represent the level of development in each platforms [4]. They remain and occupy an advanced position among the most developed technological and scientific project/research, it is important to maintain the state of their infrastructures in high quality condition/consequency to improve their performance, sustainability, and reliability [5]. Smart Platform importance of reliable infrastructures in user’s life, as well as the enormous costs associated with rebuilding infrastructures, require the deployment and implementation of predictive maintenance programs based on intelligent software techniques, accompanied by periodic and corrective maintenance where needed [6-8]. Nevertheless, it must be considered that the scale of these infrastructures is very large and that the conventional methods of inspecting infrastructures are very fast and time-consuming. These conventional methods are also dangerous for inspection team since the inspectors need to move or even climb on massive infrastructures to inspect places and areas that are difficult to reach[9-11]. Besides, most conventional inspection methods are visual and the approaches applied are manual [12]. Because of these problems mentioned above, adaptive smart platform and intelligent software have become important issues for users-aspect, industry-side, and Nation-oriented as time goes on. Eventually, we describe the ‘A’, ‘S’, ‘P’. 1) A; Adaptation mean that platform learn by itself what users and software is operating as well as to operate in the designed environment without changing their system. 2) S; Smart means software and hardware with knowledge and intelligence. In addition, smart is a digital device that allows users to expand and reconfigure functions. P; Platform is a technology that enables product development or a process that supports current or future development. And also, this is a common execution environment that provides various functions. Therefore, adaptive smart platform is adaptation technologies that are used as based upon with intelligent SW (applications), processes or basic hardware-oriented, as well as use specific machine learning-based language with cooperative different platform.

Intelligent software is an autonomous process that performs tasks on behalf of the user for a specific purpose. This is a system that does not exist independently and is part of an environment or operates within it. The environment here refers to an operating system, a network, and so on [13]. Intelligent software has a knowledge-based and reasoning function, and attempts to solve problems through information exchange and communication with users, or resources. Intelligent software has the ability to recognize changes in the environment by itself, take action to respond, and learn based on experience, and they don’t only perform a given task passively, but has an active attitude to pursue its purpose with its own purpose [14]. Intelligent Software operates inside an adaptive smart platform and provides the following features [15]. 1) Autonomy. It operates by judging by itself without direct interference from people or other objects, and has some kind of control over their behavior or internal state [16]. 2) Social Ability. It can interact with user and other software component using a communication language [17]. 3) Reactivity. It recognizes the real world, users, such as graphical user interface, a set of different software component, and the Internet-like environment and responds appropriately in time to the changes that take place within them [18]. 4) Proactivity. This is not simply acting in response to the environment, but taking initiative and acting goal-oriented [19]. 5) Temporal continuity. This is not a process that simply processes a given input and displays the result and exits, but rather a continuous daemon-like process that runs in the front and pauses for a while on the back [20]. 6) Goal-driven. It performs complex high-level tasks. The work is divided into smaller detailed tasks, and the processing order is determined and processed [21-23].

Big Data-based Adaptive Smart Platform(ASP) provide highest technology platform to helped solve a various type data problems. Such problems involve organizing miscellaneous, high volume, and potentially messy data sources to refine timely and actionable. This systems enabled new solutions to established data problems, such as search platform, social networks-based, and cloud computing and data-driven service providers. Recently, Big Data-based Adaptive Smart Platform also contribute to first line, business decisions making of industrial technology, including stock, finance, manufacturing, media, and so on. Big data platform perform data collection/indexing/retrieval, data pattern matching and abnormally detection, formal natural language processing for domain identification and semantic analysis, information extraction from abnormal data(image, audio, video data, or a variety of other data). We are shown characteristics of big data-based adaptive smart platform include as bellows:

1) Storage capacity; It collects various information from the big data environment.

2) Distributed components; It processes software component in massive connected by social network environment.

3) Data Classification; It classifies the variety of data from big data layers.

Eventually, Big platform technology is high-3V(Volume, Velocity, Variety) information, as well as it is important to enhance the capability of their systems. For big data-based adaptive smart platform, our approach provide three conceptual components: Data access patterns, Computation patterns, adaptation patterns, and we shown by <Figure 1>.

From the point of view of big data analytics, To take advantage of the multi processing that Hadoop provides, it needs to express data query with a MapReduce such as ‘text analytics – social network feeds, emails, messenger, blogs, facebook, and instagram so on’. MapReduce process by breaking the processing with map phase and reduce phase, and each phase has key-value pairs of which specifies two functions; map function and reduce function. Map function is only a data preparation phase, setting up their data, and also a good place of to drop bad records. <Figure 2> shows MapReduce logical data flow.

As we mentioned above, MapReduce Job is a unit of operate that client-side wants to be performed. This consists of the input data (cat*), MapRduce, and configuration information. Hadoop runs the job by dividing into tasks (map tasks and reduce tasks). There are two types of nodes that control the job execution with both jobtraker and tasktracker. Jobtraker run on the big data system by reports to keeps a record of the overall progress of each job using tasktracker. After all, Hadoop divides the input to MapReduce job into input splits, it creates one map task for each input splits, which runs the user defined map function for each record in their split. Splits is meant to the time taken to process, there is small compared to the time to process the whole input with by logical data flow in <Figure 2>. In order to most jobs, a good spilt size tends to the size of an HDFS block, although this could be changed for the cluster or specified when each file is created. So, we called data locality optimization, Hadoop is running the map task on a node or doing their best the map. Even though, HDFS block don’t use valuable cluster bandwidth, but input split of map task is trying to run other map tasks, so the job scheduler will find a free map slot on a node in the same blocks. we are illustrated in <Figure 3>

When we have proposed data flow by each nodes for the general case, multiple reduce tasks is shown in <Figure 4>, and this diagram makes it clear how each data flow between map and reduce tasks. Futhermore, shuffle in <Figure 2> is more complicated that this diagram suggests, and tuning it can have a job execution time.

Specifying a combine function is defined using the MapReduce class for this big data-based adaptive smart platform. It is implementation as the MapReduce Job function using Java as bellows. This Job object forms the specification of the job and gives control over how the job is run. When we run this job on a Hadoop cluster, we will package the code which Hadoop will distribute around the cluster (JAR file). In other words, Job’s setJarByClass() method use to locate the relevant JAR file by looking for the this file containing this class. Having constructed a Job object, we specify the input and output paths. After all, input path is specified by calling the static addInputPath() method on FileInputFormat, and it can be a single file, a directory, or a file pattern. Output path is also specified by the static setOutputPath() method on FileOutputFormat, that it means a directory where the output files from the reducer functions are written by Figure 3 and 4.

　public class MaxtextWithCombiner {

　　public static void main(String[] args) ... {

　　　if (args.length !=2) {

　　　system.err.println(“usage: MaxtextWithCombiner 〈input

　　　path>” + “〈output path〉”);

　　　system.exit(-1);

　　　}

　　　Job job = new Job ();

　　　job.setJarByClass(MaxtextWithCombiner.class);

　　　job.setJobName(“Max Text”)

　　　FileIputFormat.addInputPath(job, new path(args[0]));

　　　FileOutputFormat.addInputPath(job, new path(args[0]));

　　　Job.setMapperClass (..........);

　　　Job.setCombinerClass (........);

　　　Job.setReducerClass (.......);

　　　} ,..... }

In this paper, we have proposed adaptive smart platform for intelligent software in big data environment. The main idea of this research paper is to how we could have a concept big data. Other point a view, 3V of big data primary focus has been on analytics to obtain valid/valuable from adaptive smart platform. We also deal with Hadoop, HDFS, and their map flow. After all, big data platform, so we generally called ‘Adaptive platform’, ‘intelligent platform’, or ‘smart platform’, makes the case for new paradigm of their environment, as well as they have statistical methods manage for larger data sets. Technological advances of computations have enable data-driven, data classification, data optimization by using data mining, machine learning, deep learning for big data in timely manner. Especially, predictive analytics comprise a variety of big data techniques that predictive future based on historical/current data among their analytics. In addition, this analytics primarily based on statistical methods, so called Python and R, are developing new statistical methods for big data. Accordingly, our conceptual in adaptive smart platform derive on statistical significance the challenges of computational efficiency as following characteristics; Heterogeneity, Noise Accumulation, Spurious Correlation, Explanatory Data Variables, Predictor. Accordingly, from now to near future, we try to approach to integrate real data of industrial platform, in which the use of a informal data resource discovery, and manage the adaptation of smart-*, intelligent-*, and super connectivity world.