Federated Learning Model for Auto Insurance Rate Setting Based on Tweedie Distribution

In the assessment of car insurance claims,the claim rate for car insurance presents a highly skewed probability distribution,which is typically modeled using Tweedie distribution.The traditional approach to obtaining the Tweedie regression model involves training on a centralized dataset,when the data is provided by multiple parties,training a privacy-preserving Tweedie regression model without exchanging raw data becomes a challenge.To address this issue,this study introduces a novel vertical federated learning-based Tweedie regression algorithm for multi-party auto insurance rate setting in data silos.The algorithm can keep sensitive data locally and uses privacy-preserving techniques to achieve intersection operations between the two parties holding the data.After determining which entities are shared,the participants train the model locally using the shared entity data to obtain the local generalized linear model intermediate parameters.The homomorphic encryption algorithms are introduced to interact with and update the model intermediate parameters to collaboratively complete the joint training of the car insurance rate-setting model.Performance tests on two publicly available datasets show that the proposed federated Tweedie regression algorithm can effectively generate Tweedie regression models that leverage the value of data fromboth partieswithout exchanging data.The assessment results of the scheme approach those of the Tweedie regressionmodel learned fromcentralized data,and outperformthe Tweedie regressionmodel learned independently by a single party.

Dyeing Thermodynamics and Supramolecular Structure of Lac Red on Protein Fibers

The dyeing temperature of natural dye lac red on two kinds of natural protein fibers was studied, and the interaction between dyestuff and fiber was discussed through thermodynamic study and density functional theory (DFT) calculation. The optimum temperature for lac red dyed silk was 60˚C and wool showed a better response at 90˚C. The thermodynamics study revealed good Nernst isotherm and Freundlich adsorption models respectively, and the lac dye adsorption processes were both spontaneous and exothermic. The potential interaction of Laccaic acid A with the external environment by electrostatic potential and atomic charge distribution was first explored. With molecular simulation, Laccaic acid A and glycine composed 8 stable complexes. Then, typical hydrogen bonds, bond length, and binding energy, etc. were analyzed. The results revealed lac red on silk and wool fabric mainly depended on the weak hydrogen bonds and van der Waals force which determined the low dye fastness.

Research of Dyeing Thermodynamics and Supramolecular Structure of Luteolin on Wool Fabric

Natural dyestuff of luteolin was isolated and used to dye wool fabric in this paper. Ethanol extraction and high-speed countercurrent chromatography (HSCCC) were used to extract and purify the luteolin from the peanut shell, and the structure of the isolated luteolin was characterized with FTIR techniques. The interaction between dyestuff and fiber was preliminarily discussed through thermodynamic study and supramolecular structure simulation to explain the intrinsic reasons why the color fastness was low when luteolin was applied to dyeing wool fabric. The extraction condition and purification parameter were as follows: 65% ethanol, ratio of material to liquid 1:20, 80°C, 3 h, chloroform-methanol-water (4/3/2, V/V), 800 rmp/min, 2.0 Mkpa, 0.5 mL/ min and 280 nm. The results of dyeing thermodynamics showed that the sorption isotherm of luteolin on wool fabric was consistent with Nernst model and similar to the disperse dyestuff. With molecular simulation, luteolin and glycin composed 8 stable complexes whose Laplacian values all were greater than 0, which suggested typical hydrogen bonds existing. The complex with three hydrogen bonds was proved the most stable. Both studies on thermodynamics and supramolecular simulation revealed that luteolin on wool fabric mainly depended on the weak hydrogen bonds interaction that determined the low dyefastness.