With the strengthening of domestic Internet financial supervision, the position of small-scale inclusiveness has become more clear (the nature of industrial regression risk control). It has become more and more important to use big data and machine learning as a means of risk control, and companies in related fields are gradually receiving Sought after. Lei Feng Network was the only one to learn that Fintech CreditX recently completed tens of millions of pre-A rounds of financing, and was led by the newfound fund of volcanic stone funded by Zhang Suyang, the IDG's honorary partner. At the end of 2015, Ruan Xin also received 7 million yuan angel financing led by a real estate fund. Yuxin mainly uses machine learning and big data analysis technologies to provide risk credit assessment and user value operations, providing financial companies with risk control and marketing decisions. Lucent's products include non-random engines and XCloud, the former is a big data risk control decision engine; the latter can be simply understood as DAAS (data as a service) services based on big data. Company founder Zhu Mingjie told Lei Fengnet that this round of financing will be mainly used for upstream and downstream cooperation with data sources, financial scenarios and market expansion, and team expansion. Internet lending is characterized by a large amount of money, while artificially-based risk control is costly and inefficient. This requires machine learning to help make risk control decisions. Zhu Mingjie believes that it is still at an early stage to use machine learning as a method of risk control. For financial customers, there is also a market education process. This year, only gradually began to have business needs. He told Lei Fengwang (searching for "Lei Feng Net" public concern) that the application of machine learning to risk control has two main problems: One is too much data, mainly data dimension. Traditional risk control mainly uses financial-related features. There are generally dozens of dimensions, including income, running water, and credit. Small inclusive finance requires multidimensional data including behavioral data, APP data, and long-term borrowing. Second, the sample data is too small, because the pure line form such as consumer finance exists for a short time, and there is no sample data as much as a traditional loan. The Trusted Machine Learning Engine will combine human experience and machine capabilities. In the online loan, although the data is insufficient, the risk control expert will have his own logic, but the amount of manpower review will not be great. The experience is copied to the machine through machine learning, and the uncertainty is given to the risk control expert. This is suitable for small-scattered loans. The risk control expert will intervene in the training iterations of machine learning. The final model will be automatically iterated, and the dependence on people will be less and less, and the moral hazard of human operations will be avoided. Zhu Mingjie also mentioned the use of machine learning for interpretable issues of wind control. We know that although technologies such as deep learning have achieved amazing results in face recognition and speech recognition, there are still few discussions about their principles, that is, they do not know why such results have been achieved. Financial institutions are particularly concerned about the interpretability of the model. On the one hand, it is necessary to control risks. On the other hand, it is necessary to communicate with users. Obviously, it is unacceptable to give results without explanation. Lei Fengnet has reported that Lixin's experience is to use LIME to capture the key variables in the results or local results, and then let the risk control experts quickly grasp which features lead to changes in results. Zhu Mingjie stated that since the product was formed at the end of last year, Dangxin has cooperated with more than 50 financial institutions, including traditional financial companies such as Minsheng Bank and Bank of China Consumer Finance, and it is also a little financial, financial, financial, and rain. Jinfu and other internet finance companies. Non-random engine In the traditional financial scenario, the KS value of the risk control indicator is raised above 50%, and the KS value is also leading in the anti-fraud effect of Internet finance. At present, the company has achieved profitability and its financing is mainly for the future layout and platform development. Zhu Mingjie said that the interval between the two rounds of financing is relatively short, and it is also very soon that Zhang Suyang has retired from IDG to form a new fund. Zhu Mingjie told Lei Fengwang that many of the previous big data and machine learning challenges were storytelling. With the tightened supervision , the Fintech infrastructure service has only just begun. The underlying infrastructure includes data, models, Credit information, risk control and anti-fraud. However, he believes that this market will have a significant Matthew effect. Although more companies will join it in the future, the start-up team will have a greater advantage. Because they have test results in mature scenes, the model will be more and more perfect. There are several barriers to doing wind control, one is talent, and the other is the accumulation of data and business. The machine learning model needs to be tested by financial scenarios. The effect is a combination of technology and business. Zhu Mingjie believes that the advantage of Yuxin is also here. He is a Ph.D. from Microsoft Research Asia. He was engaged in large-scale data mining at the Max Planck Institute in Germany. Other technical members also have mature large-scale machine learning mature application experience. In addition, they also have a very good opportunity to go deep into the scene of risk control at the core of financial institutions so that products can be tested. At present, Lucent's risk control engine mainly provides scene-customized services, but it has already cooperated with cloud financial platforms such as Sina Wealth , and will provide more financial SaaS services in the future, so that the wind control model can be effectively promoted and replicated. High-temperature Submersible Motor Winding Wire
This products is made of solid copper wire or enamelled copper wires conductor, crossed-linked polyethylene insulation, optional polyamide jacket. It has fine performance for water-resistant, electric conducting and mechanical. It used to all types water-filled, oil-filled high-temperature submersible motor or similar working electronics, and applied to standard of environmental-friendly.
The conductor is used the standard of GB/T395 or GB/T6109, and applied to JB/T4014 standard.
The product can be used for Voltage 1140V, temperature -40℃-90℃, PH value 6.5-8.5, water pressure 1MPa, light acid-alkali-salt media or oil-media, frequency 50Hz-60Hz, for long time use with stable liability.
High-Temperature Submersible Motor Winding Wire,Frequency Conversion Submersible Motor Winding Wire,Frequency Covertable Submersible Winding Wire,Heat Resistance Submersible Motor Winding Wire HEBEI GUOQIAN WIRES CO., LTD , https://www.guoqianwires.com
Using Machine Learning for Financial Risk Control
Big story data control that no longer tells stories
