1. Report Industry Investment Rating - Not provided in the content 2. Core Viewpoints of the Report - The report introduces the LSTM time - series model to improve the prediction of Treasury bond futures inter - period spreads without expanding the original factor set, finding that introducing time - series modeling can improve return stability and risk control for some varieties, but the sensitivity to time - series information varies among different contracts [38][39][40] - The LSTM model outperforms the DNN model on T and TF contracts, while the DNN model shows stronger prediction ability on TS contracts; the performance of both models on the TL contract is limited due to the short listing time [38][39] - Using the LSTM model's prediction probability for position weighting can enhance the return - risk characteristics of strategies on T, TF, and TS contracts, but the effect is not obvious on the TL contract [38][39] 3. Summary by Relevant Catalogs 3.1 Inter - period Spread Influencing Factors and Indicator Selection - Traditional machine - learning models and DNNs have limitations in time modeling as they assume sample independence in the time dimension and rely on manual feature engineering to introduce time information, and DNNs lack explicit sequence structure [7] - LSTM is an improved form of RNN, introducing a memory unit to decouple long - term information storage and current state output, and using gate - control structures to control information flow, which can better capture long - term dependencies [7] - Treasury bond futures inter - period spreads show trend characteristics in some stages, which are difficult to capture by previous models. Therefore, the LSTM time - series model is introduced to enhance the description of the time - series structure [12] 3.2 Recurrent Neural Network Testing Process 3.2.1 Data Processing and Sample Construction Process - The data includes fundamental factors of T, TF, TS, and TL contracts, and the features are constructed by aligning and introducing capital - related indicators and creating derived spread variables. The label is the first - order difference of the inter - period spread [16][17] - Data pre - processing involves removing early - listing samples and some end - of - month trading days, filtering small - amplitude spread changes, filling missing values, and standardizing data [18] - Time - series samples are constructed using a sliding window with a 5 - trading - day historical window, and samples are divided into training, validation, and test sets by strict time segmentation [19][20] - The model is trained using a weighted cross - entropy loss function and the Adam optimizer, with learning - rate decay and early - stopping mechanisms based on validation - set loss [22] 3.2.2 Parameter Setting and LSTM Network Structure - The task is a binary - classification prediction for the next - trading - day direction change of the inter - period spread, with input as a sequence of factor values over 5 consecutive trading days and output as binary logits [24] - The LSTM network has 3 layers, a hidden - state dimension of 8, and a Dropout ratio of 0.3. It uses the hidden state at the last time - step for classification [25] - The weighted cross - entropy loss is used to address class imbalance, and the Adam optimizer with learning - rate decay is applied for parameter updates [26][27] 3.3 Model Test Results 3.3.1 Comparison of Out - of - Sample Tests between LSTM and DNN Models - On T and TF contracts, the LSTM model has higher cumulative returns, Sharpe ratios, and better drawdown control compared to the DNN model; on the TS contract, the DNN model performs better; on the TL contract, the performance of both models is limited [28][30] 3.3.2 Probability - Weighted Backtesting of the LSTM Model - The prediction probability of the LSTM model is used as the position signal, and the position is normalized and limited to control risks [33] - Probability - weighted strategies improve the out - of - sample performance on most contracts, enhancing returns and Sharpe ratios without significantly increasing drawdowns, but the effect is not obvious on the TL contract [35] 3.4 Conclusion - Introducing the LSTM time - series model can improve the prediction of Treasury bond futures inter - period spreads for some varieties, but the effectiveness depends on contract characteristics, sample coverage, and spread structure [38][39][40] - Using prediction probability for position weighting has potential in improving strategy performance, and future research can further test the applicability of time - series modeling under more complex conditions [39][40]

Report Industry Investment Rating No relevant content provided. Core Viewpoints of the Report - The sentiment in the bond market has shifted from "bullish but not buying" to a stage where some bulls are "wavering." The short - term trading themes are broad - based monetary policy and the fundamentals. The 10 - year Treasury bond is oscillating around 1.65% - 1.70%, and the bond market is likely to be range - bound in the short term. The impact of positive events on the bond market is rapid this year. The proportion of capital gain demand in the comprehensive return has been continuously increasing, so the duration should be maintained, and leverage can be appropriately increased if the funding rate declines [2]. - From the perspective of market sentiment, the bond market has changed from "bullish but not buying" to a stage where some bulls are "wavering." This week, some bulls have turned neutral, the number of institutions with bearish views has increased by one, and the sentiment index has declined [3]. - The fundamentals and broad - based monetary policy are the "battlefields" for the bulls and bears among the sellers. As of May 19, the number of fixed - income sellers with bullish views has decreased to 10, the number of those with bearish views has increased to 3, and the number of those with neutral views has risen to 17 [3]. - Among the buyers, those with neutral views also account for more than half, and the proportion of institutions with bullish views has decreased. The overall view of fixed - income buyers is neutral - bullish. Currently, there are 10 bullish, 16 neutral, and 3 bearish institutions [3]. - In the Treasury bond futures market, the positive arbitrage space for the TS contract has decreased, and it may still be in a premium state. One can consider participating in the game of the TS contract rising [6]. Summary According to Relevant Catalogs 1. Seller and Buyer Markets 1.1 Seller Market Sentiment Index and Interest - rate Bonds - The seller sentiment index has decreased compared to last week. The weighted index this week is 0.18 (neutral - bullish, down 0.10 from last week), and the unweighted index is 0.28 (down 0.14 from last week). Currently, institutions generally hold a neutral - bullish view, with 10 bullish, 17 neutral, and 3 bearish. 33% of institutions are bullish, with keywords such as the long - term trend of Sino - US decoupling despite tariff fluctuations, a possible LPR cut this week, and the expected continuation of monetary easing; 57% are neutral, with keywords such as high macro - environment uncertainty, waiting for a new market trigger after the double - cut and tariff easing, and the bond market may have a narrow - range oscillation; 10% are bearish, with keywords such as the exhaustion of the double - cut benefits, the lack of support for the bond market reflected by the non - rising funding rate, and the central bank may take measures to maintain the bank's net interest margin and push up the long - term bond yield after the tariff cut [13]. 1.2 Buyer Market Sentiment Index and Interest - rate Bonds - The buyer sentiment index has decreased compared to last week. This week's sentiment index is 0.18 (neutral - bullish, down 0.22 from last week). Currently, institutions generally hold a neutral - bullish view, with 10 bullish, 16 neutral, and 3 bearish. 35% of institutions are bullish, with keywords such as continuous loosening of the funding side, the economy still needing policy support, and the reduction of funding costs; 55% are neutral, with keywords such as fluctuations in Sino - US economic and trade expectations, frequent policy disturbances, unclear fundamental expectations, differentiated interest - rate trends, limited adjustment space but repeated directions; 10% are bearish, with keywords such as the marginal weakening of easing expectations, the enhanced expectation of economic fundamental repair, frequent funding disturbances, and increased long - end supply pressure [14]. 1.3 Credit Bonds - Market hot topics include policies to promote science - and - technology innovation bonds and implicit debt accountability. Multiple departments have introduced policies to promote the construction of science - and - technology innovation bonds, and it is expected that future issuance increments will be for financial institutions and private enterprises, covering more science - and - technology innovation fields. The Ministry of Finance has emphasized local government implicit debt governance again, and the issuance supervision of urban investment bonds has become stricter, with risks being relatively controllable in the short term [18]. 1.4 Convertible Bonds - This week, institutions generally hold a neutral - bullish view, with 2 bullish and 6 neutral. 25% of institutions are bullish, with keywords such as the positive impact of the double - cut and the end of the earnings disclosure period, the increased risk appetite in the convertible bond market, and the strong equity market becoming an important support for convertible bonds; 75% are neutral, with keywords such as the current high valuation of convertible bonds, low cost - effectiveness, a possible range - bound oscillation pattern, and the need for incremental funds or overall underlying stock repair for a stronger market [20]. 2. Treasury Bond Futures Tracking 2.1 Futures Trading - In terms of futures prices, except for the increase in the TS contract price, the prices of other futures contracts have decreased. As of May 16, the prices of the Treasury TS/TF/T/TL contracts are 102.38 yuan, 105.72 yuan, 108.48 yuan, and 118.91 yuan respectively, with changes of +0.04 yuan, - 0.38 yuan, - 0.58 yuan, and - 1.46 yuan compared to last Friday. - In terms of Treasury bond futures open interest, except for the increase in the TS contract open interest, the open interest of other futures contracts has decreased. As of May 16, the open interest of the TS/TF/T/TL futures contracts is 84,000 lots, 79,000 lots, 100,000 lots, and 50,000 lots respectively, with changes of +702 lots, - 64,061 lots, - 76,980 lots, and - 31,940 lots compared to last Friday. - The trading volume of Treasury bond futures has increased across the board. As of May 16, from a 5 - day moving average perspective, the trading volumes of the TS/TF/T/TL futures contracts are 132.9 billion yuan, 98.5 billion yuan, 131.1 billion yuan, and 155.6 billion yuan respectively, with increases of 40.2 billion yuan, 25.9 billion yuan, 44.7 billion yuan, and 39.9 billion yuan compared to last Friday. - The trading volume - to - open - interest ratio of Treasury bond futures has increased across the board. As of May 16, from a 5 - day moving average perspective, the trading volume - to - open - interest ratios of the TS/TF/T/TL futures contracts are 1.01, 0.98, 1.03, and 2.58 respectively, with increases of 0.49, 0.49, 0.56, and 1.33 compared to last Friday [24][25]. 2.2 Spot Bond Trading - The turnover rate of 30 - year Treasury bonds has decreased. On May 16, the turnover rate was 2.32%, down 0.10 percentage points from last week and 0.41 percentage points from Monday, with a weekly average turnover rate of 3.33%. The turnover rate of interest - rate bonds has decreased. On May 16, the turnover rate was 0.89%, down 0.05 percentage points from last week and 0.22 percentage points from Monday. The turnover rate of 10 - year China Development Bank bonds has increased. On May 16, the turnover rate was 5.81%, up 0.17 percentage points from last week but down 1.54 percentage points from Monday [35][36]. 2.3 Basis Trading - In terms of basis trends in the past week, the basis of the TF main contract has narrowed, while the basis of other main contracts has widened. As of May 16, the basis (CTD) of the TS/TF/T/TL main contracts is - 0.07 yuan, +0.05 yuan, +0.11 yuan, and +0.17 yuan respectively, with changes of - 0.03 yuan, +0.10 yuan, +0.17 yuan, and +0.05 yuan compared to last Friday. - In terms of net basis, the net basis of the TS main contract has widened, while the net basis of other main contracts has narrowed. As of May 16, the net basis (CTD) of the TS/TF/T/TL main contracts is - 0.05 yuan, +0.03 yuan, +0.02 yuan, and +0.03 yuan respectively, with changes of - 0.06 yuan, +0.12 yuan, +0.08 yuan, and +0.09 yuan compared to last Friday. - In terms of IRR, the IRR of the TS contract has increased, while the IRR of other main contracts has decreased. As of May 16, the IRR (CTD) of the TS/TF/T/TL main contracts is 1.79%, 1.32%, 1.46%, and 1.39% respectively, with changes of +0.35%, - 1.02%, - 0.66%, and - 0.56% compared to last Friday. The TS main contract's basis is negative this week, and the weekly average IRR is 1.65%, at a relatively high level. Since the funding side is generally in a stage of loosening this week, with the weekly average DR007 at 1.54%, one can pay attention to the positive arbitrage strategy of the TS contract [41][44][45]. 2.4 Inter - delivery Spread and Inter - product Spread - In terms of inter - delivery spread, the spread of the T contract has widened, while the spreads of other main futures contracts have narrowed. As of May 16, the near - month minus far - month spreads of the TS/TF/T/TL contracts are - 0.11 yuan, - 0.20 yuan, - 0.17 yuan, and - 0.34 yuan respectively, with changes of +0.09 yuan, +0.14 yuan, - 0.00 yuan, and +0.19 yuan compared to last Friday. - In terms of inter - product spread, the spreads of the 2*TS - TF and 4*TS - T contracts have widened, while the spreads of other main futures contracts have narrowed. As of May 16, the values of 2*TS - TF, 2*TF - T, 4*TS - T, and 3*T - TL are 99.04 yuan, 102.93 yuan, 301.01 yuan, and 206.50 yuan respectively, with changes of +0.43 yuan, - 0.25 yuan, +0.62 yuan, and - 0.29 yuan compared to last Friday. Currently, the downward space for long - term interest rates is limited. If the central bank takes measures to ease liquidity, there may be downward opportunities for the medium - and short - term. There is considerable room for gaming in short - term Treasury bond futures, and it is recommended to continue to pay attention to the strategy of going long on the short - end and short on the long - end to steepen the yield curve [51][52].