1. Company introduction MTOV Co., Ltd. conducts various missions in automotive-related areas and innovates smart driving. 2. Problem Background and Summary OBD is a legally prescribed sub-diagnosis/monitoring system integrated into the organ control system. Using OBD data, we want to model safe driving indices that cannot be done on navigation or that use two. The goal is to calculate and model indicators (scores/classes that can measure the degree of safe driving) that can help individual drivers or insurance companies provide reasonable evidence of insurance premiums (discounts/ surcharges or subion/rejects). It is necessary to model a more reasonable and reliable safe driving index, not to calculate scores from data such as rapid acceleration, rapid acceleration, and number of idling.     3. Solving Process Calculation of safe operation figures using operation/shifting by speed section Proposal of a method for calculating speeding and slow driving figures using speed/limit/average speed (flow of traffic flow) Proposal of Weight by Accident in Operating Section/Time Zone

1. Company introduction 2. Problem Background and Summary US real estate company Zillow provides services that predict how the housing prices they own will change in a year. DH Engineering & Construction Co., Ltd. aims to benchmark Zillow and "develop a model for predicting housing value based on mathematical algorithms" Analysis of Methodology Revealed by Zillow, Modification of Zillow Methodology to suit the Korean Housing Value Prediction Model, Analysis of Housing Status Required by Model, Propose factors that affect local values 3. Solving Process Since Korea is heavily influenced by prices according to real estate policies, the Zillow methodology has been modified so that the sensitivity of the housing value prediction model is not measured in the past year, but in the past three years, so that it does not fluctuate much in the impact of the real estate boom or bad news. An Analysis of the Correlation between the Factors Affecting Housing Prices in Gangnam-gu District. A Study on the Housing Price Determinants Using the Hedonic Price Model : Focused on the Ichon-dong Area, Seoul, by referring to the Ministry of Land, Infrastructure and Transport's actual transaction price disclosure system and Lee Kang and Choi Shin-hee's paper The research was conducted in Kim Joong-pyo and Hong Sung-jin's paper "Research on Housing Price Determinants in the Center of Large Cities: Focusing on Jung-gu, Daegu" and proposed data elements that can be obtained through Google API such as fire stations, police stations, kindergartens, daycare centers and hospitals closest to housing.   4. Ripple effects and future plans It is expected that the systematic collection of data on proposed factors to measure housing status and local value will enable the verification of the methodology for the Korean model. The methodology presented through model verification needs to be modified and supplemented. Using the proposed model, we expect the value of the company to increase if DH Construction Co., Ltd. provides housing value forecasting services first.

1. Company introduction MTOV Co., Ltd. conducts various missions in automotive-related areas and innovates smart driving. 2. Problem Background and Summary It is important to measure accurate fuel consumption and fuel economy for economic driving or autonomous driving. The fuel economy measurement method using mileage and fuel meter residual fuel volume can only find the average fuel economy in a relatively long section, and depending on the condition such as the inclination of the vehicle, the remaining fuel volume measurement sensor also has errors, making it difficult to accurately measure. The goal is to identify existing fuel economy measurement models to derive accurate fuel economy measurement methods, or to model new measurement methods using OBD data.   3. Solving Process Provides basic data analysis methods and analysis results, including exploratory data analysis, statistical analysis of correlation, and visualization of OBD data By smoothing (noise removal) the remaining fuel volume of the fuel system with large errors using linear return, moving average, etc., the remaining fuel quantity provided by OBD is not detailed in the minimum unit and the fuel efficiency of the section is difficult to measure even with a large error smoothing. Implementing an instant fuel economy measurement program using OBD items such as Mass air flow (MAF) and Air/fuel ratio If OBD data is defective depending on vehicle type, model year, vehicle condition, etc., suggest and implement methods to predict items that are largely related to fuel economy calculations such as MAF with other data.   4. Ripple effects and future plans Expect to be able to more accurately calculate the actual fuel economy of the car and give the driver accurate indicators After additional and ongoing data collection by the enterprise, results can be applied to improve algorithms

1. Company introduction Define Co., Ltd. is a company that develops and supplies system software, develops and supplies optimal collection systems for household waste, eradicates harmful tides, and smart parking systems based on deep learning. 2. Problem Background and Summary Although the service is being developed by applying a deep learning model that enables real-time object detection, performance delays are involved when performing image analysis on low-performance devices, and mathematical principles and improvement directions are derived to receive suggestions on how to change or lighten the model. The goal is to develop algorithms to detect harmful tides in low-performance devices.   3. Solving Process Change settings for low-performance devices (Lotte Panda) It has been confirmed that the actual detection speed can be improved without the need to modify or relearn the current model using input dimension changes and multi-core environment of the instrument, as well as the ability to reduce latency by adjusting the number of frames. Currently in preparation for field testing by applying this method directly to the prototype, Using SSD MOBILENET V1 Models Using SSD MOBILENET, one of the most well-known deep learning techniques related to object detection, we have confirmed that the frame processing speed is higher than that of the existing use model, and we expect to get improved results from the actual low-performance device.   4. Ripple effects and future plans If excellence is proven in the field, it is expected that economic wealth will be created not only for companies but also for the nation through expansion of regions across the country, expansion of overseas markets through diversification of products, and promotion of projects to link harmful tidal waves.

1. Company introduction SmartSocial Co., Ltd. was founded in 2012 and is taking the lead in solving social problems by utilizing big data. Major business areas include resolution of job mismatching, public works, youth employee tomorrow deduction, small business exploration project, youth tomorrow morning deduction operation support, student employment management, etc. 2. Problem Background and Summary This industrial problem was discovered while commercializing matching algorithms developed by Smart Social and the Industrial Mathematical Innovation Center, and the "scoring algorithm" currently used for matching job seekers with companies does not reflect various information about job experience of job seekers, resulting in a rush of matching algorithms. The goal is to more accurately quantify individual job competency through modeling that reflects various information (period, level, etc.) about job experience of job seekers.   3. Solving Process Using 'Matching Score Correction Model A', the work-seeker's corporate matching score is corrected by redefining the job competency by reflecting the level and duration of the job experience. This model has allowed job seekers with different job levels and periods to differentiate their matching scores, and some resolution of matching congestion. 'Matching Score Corrective Model B' is a model that redefines job competency by reflecting the period after the work-seeker's experience is cut off, and the company's matching score varies depending on the period of job-experience break. This model is used in conjunction with 'Matching Score Calibration Model A' to address matching congestion 'Matching Score Calibration Model' allows you to achieve cumulative corrected scores for one or more portfolios of job seekers The career calculation model utilizes the job classification and level of the National Competency Standards (NCS), and confirms the correction effect through data from companies and job seekers who have been crowded.   4. Ripple effects and future plans It is expected that the new algorithm will be effective in enhancing the reliability and validity of the service by allowing users to explain the matching service realistically and logically by utilizing scores calculated based on individual experience. Currently, corporate matching services have been used to match students with practical companies, and the service can be presented by expanding the user base to include those in office and career-setting women.

1. Company introduction Seoul Metro, which used to operate subway lines 1 to 4, and Seoul Metropolitan Rapid Transit Corporation, which operated lines 5 to 8, were established by integrating them. 2. Problem Background and Summary The industrial problem is to improve the Smart Automatic Mechanical Big Data Analysis-system (SAMBA) established by the Seoul Transportation Corporation. The goal is to enhance algorithms that automatically recognize mechanical failure situations or predict failures through a mathematical approach. Develop a way to efficiently utilize data stored in real-time on big data servers Development of mathematical algorithms to detect abnormalities in facilities such as V-belt and motor bases and predict failures 3. Solving Process Proposal of data collection methods for detecting abnormalities in air conditioners, such as V-belt slip and departure: Securing data (abnormal data) in case of a failure is essential to conducting guidance studies such as deep learning or evaluating the performance of a model. We decided to select specific air conditioners to develop abnormal detection algorithms and expand them. Collect abnormal data by simulating failure of 10 air conditioners in Janghanpyeong Station. Visualize and interpret sensing data using time-series data synchronization techniques: To develop a model that detects anomalies, an algorithm is applied to divide time series data in one dimension into data preprocessing based on the point at which the air conditioner motor operates. Suggest synchronization method with correlation, Show how to sync with maximum values, Visualize and interpret data   4. Ripple effects and future plans If applied to the Seoul Metro's automated big data analysis system, it is expected to provide practical convenience for citizens to use the subway. We are planning to develop abnormal detection algorithms that can be applied to various air conditioners in common, not to individual air conditioners only. Anomaly detection model will be developed using vibration data, including a comparative analysis of current sensing data and vibration data.

1. Company introduction Aiming to improve the production efficiency of manufacturing factories through the construction of smart factories, the startup, which started operations in 2016, stores the data collected in various manufacturing processes on a big data platform, analyzes it, and detects product defects. Develop "Quality Guard", a solution for detecting abnormalities in production machinery. 2. Problem Background and Summary There are the following problems and improvements in the defect detection algorithm that distinguishes normal products from defective products by using data generated in the computer numerical control (CNC) manufacturing process. Limited number of types that can be detected Absence of automatic defect detection algorithm that can be applied to multiple processes Need to measure wear of tools and develop model for predicting damage. The goal is to find pretreatment methods for data analysis and to define mathematical characteristics of data and set classification criteria to clearly distinguish between normal and defective products. 3. Solving Process Define different mathematical features that can be applied to multiple processes and apply them to algorithms to improve defect coverage Research progress of synchronization technique of time series data required for data processing Developed an algorithm for setting meaningful process intervals in consideration of the cumulative amount of change   4. Ripple effects and future plans Completed support for sophistication of defect detection algorithms using basic statistics Applying the defect detection algorithm provided by the laboratory to "QualityGuard 2.0“ Scheduled patent application for research results of time-series data synchronization method Based on mutual trust during the problem-solving process, he participated in the "Start-up Growth Technology Development Project – Start-up Tasks" of the Ministry of SMEs and Startups for 17 years. ※Administrative agency: Takion Tech, Subcontracting agency: National Institute of Repair Science In 2018, the Land, Infrastructure and Transport Technology Commercialization Support Project was jointly supported by the National Transportation Safety Technology Commercialization Project and will jointly develop abnormal operation detection solutions with the aim of improving road traffic safety when selected. Lee Kwang-won, head of Takion Tech's research center, said mathematicians have a quick understanding of industrial knowledge and are efficient at solving corporate challenges by presenting creative ideas through mathematical approaches.

1. Company introduction It is a comprehensive energy service corporation that supplies city gas throughout Daejeon Metropolitan City and Gyeryong City since its establishment in 1985. ※ In October 2017, the name was changed from Chungnam Urban Gas Co., Ltd. to CNCITY Energy Co., Ltd.     2. Problem Background and Summary To ensure the safe management of city gas, a person in charge of each district conducts a daily inspection of gas pipeline burial areas and solves events (checking excavation work, handling civil complaints, etc.). Because the time and location of the event are different each time, the location of the remaining pipes to be checked after the event is always different, and the current inspection order of the remaining pipes is determined solely from the experience of the person in charge. Using mathematical principles such as graph theory, etc., the goal is to develop an algorithm to find the optimal route for the remaining piping circuit inspection, to check if the current method is the best, and to find a more efficient way of circuit inspection.       3. Solving Process An algorithm was developed to print the shortest route back to the head office when the current location and remaining pipes were entered. Utilizing the principles of Chinese Postman Problem and Minimal perfect matching in graph theory, we find the shortest path, and show this on the map.     4. Ripple effects and future plans Identifying the possibility of optimal circuit inspection routing using the developed algorithm

1. Company introduction SmartSocial Co., Ltd. was founded in 2012 and is taking the lead in solving social problems by utilizing big data. Major business areas include resolution of job mismatching, public works, youth employee tomorrow deduction, small business exploration project, youth tomorrow morning deduction operation support, student employment management, etc.   2. Problem Background and Summary Based on the National Competency Standards (NCS), a matching algorithm using the enterprise's own standards required mathematical analysis and visualization of data produced during the development and operation of the platform. The aim is to analyze the relationship between the data held, to classify student propensity and to create visualizations that can be understood by the general public. Research is needed for classifying and processing information that students write themselves or collect from schools, and for analysis of text documents (student-generated portfolios), such as checking the consistency between the content of the portfolio and the classification of NCS jobs.   3. Solving Process The institute presented methods of data calculation and mathematical visualization for job matching analysis. Presentation and analysis methods of the Recurring Genome Map (student classification using K-NN clustering) created through matching scores between students and corporate practice calculated from matching algorithms were presented. System division for smooth management according to the increase in information collected (Score Algorithms 1, 2 and Matching Algorithms division) was presented. Through natural language processing-word frequency analysis for NCS classifications analysis, prototype type development and prototype of job similarity check NCS filter were presented. 4. Ripple effects and future plans Based on the results of task matching analysis data calculation, it has ed installing a platform for visualization contents (such as Recurring Genome Map-Heatmap, Job Tree-Tree Diagram, etc.). By using NCS filter, it succeeded in attracting investment for chatbot development. It is planning to support small and medium venture business department's new product development project in 2018.  

1. Company introduction VOVIOS KOREA is a North American online order platform and mobile delivery platform management business established in 2015.     2. Problem Background and Summary We are conducting research to automate the problem of relocation of transit points due to multiple delivery scheduling without human intervention. When a new delivery order comes in, it tries to find an optimal route through relocation of transit points. While all time measurements using Google APIs are accurate, there is a need for a way to reduce API usage by increasing the number of API requests and latency due to the increase in transit points. Currently, the approximate distance traveled is calculated using the shortest straight line distance calculation method, but a request was made to improve the distance calculation method that reflects the local topography such as legs and elevated roads.     3. Solving Process For the relatively accurate calculation of travel distances, methods for calculating distances, such as Taxicab metric, which reflects topographical characteristics, rather than the shortest straight distance method, were proposed. Area division using point in polygon propulsion (PIP) was presented to minimize errors in calculating travel distance due to bridges, overpasses, etc. and to enable response by delivery area. Algorithms and codes of the recommendation calculation function were provided to provide routes tailored to the priorities of the variables desired by the enterprise. 4. Ripple effects and future plans Kang Byung-hak, CTO of VOVIOS KOREA, said, "I am surprised and grateful that there is a lab that supports start-ups that has presented all of the company's requirements." He hoped that continued research and development would take place after the platform of the proposal, but due to the circumstances of Vovios Networks Inc. in Vancouver, the domestic Bovios Korea project has stopped because it is no longer in progress.